Sudan_catch_analysis

Analysis of Sudan survey data 2012 - 2013

This R-markdown document contains all r-code used to carry out the analysis for the paper on fish catch rates and species densities (species richness) along the Red Sea coast of Sudan. The analysis is based on three surveys: November 2012, May 2013 and November 2013.

Here the analysis is structured around the 7 management regions of the Sudanese Red Sea coast as follows:

• Present the region: Bathymetric map
• Sampling locations
• Table of effort (diffent gear types per management areas.)
• Depth distribution of traps by survey and area
• Catch CPUE by area and survey
• Catch CPUE distribution by gear, by species groups (should perhaps do the same for trophic groups)
• CPUE by trophic group pr area and survey
• CPUE by depth and tropich group and survey
• Biodiversity /species densities
• Statistical analysis
• test of normality shows that we need to use non-parametric and zero-adjusted approaches.
• CPUE as a function of area, depth, survey, gear (Zero-adjusted model)

All code and data are stored on GitHub

Libraries, data etc.

Libraries, dependencies and functions

Load and manipulate data

Reading catch, station and traits data.

Neither station data, nor catch data has a complete depth record, but by combining depth para from each we can get a complete depth parameter for all stations.

Sudan map and management areas

Loads the map data for Sudan, loads the management areas from shape-file etc.

## OGR data source with driver: ESRI Shapefile 
## Source: "/Users/eriko/Documents/GitHub/Sudan2019/Sudan-master/sudan_management_areas", layer: "sudan_regions"
## with 7 features
## It has 3 fields
## Integer64 fields read as strings:  id

Allocating catch positions to management areas

Adds ‘Area’ to each line in the catch table.

Modifying dataset and creating one dataset for CPUE analysis and another for traits

Of the catch data, 12 fish registrations lack weight, (i.e. this was forgotten entered into the database during the survey). These registrations, would if included lead to 12 more registrations of CPUE = 0.

Adding traits to ‘catch’ data set (without 0-catch stations)

Some more data wrangling that adds the traits from the traits table to the catch data, using only the station with catches (there are no traits for ‘NOCATCH’ species).

Also, only select traps, gillnets and handlines as these were the only gear with sufficient numbers and consistent use to be analyzed.

Lastly adds number of gear deployed at each station to each line.

Bathymetric map of Sudan

Fig. 1 in MS

A nice, bathymetric map of Sudan with management areas overlaid.

Currently fonts and colours are adjusted for use on the ICES ASC 2019 poster, removing the bathy-contour lines (se ‘lwd’ to 0), and varying the colours of management area text to white for areas 1-4.

## quartz_off_screen 
##                 2

Stations plotted on maps, pr survey

Faceted map plotting position of all catch stations for each of the three surveys.

## Saving 7 x 5 in image

Station information

Station table

Table describing the sampling effort in each area pr survey, number of different gear types, max, min and average depth, number of traps with / without catches.

survey	id	Ntraps	Nhl	NGn	TBhrs	HLhrs	GNhrs	DepthAvg	DepthSD	DepthMax	DepthMin
2012901	1	22	0	0	694.066	0.000	0.000	42.45455	27.104256	142	13
2012901	2	54	3	3	721.549	78.000	62.000	40.89744	16.597075	71	0
2012901	3	26	0	1	451.183	0.000	14.000	30.80769	19.237503	95	8
2012901	4	5	0	0	77.084	0.000	0.000	22.60000	17.910891	54	10
2012901	5	31	0	4	677.896	0.000	78.317	21.32258	6.498139	30	7
2012901	6	36	0	1	850.417	0.000	38.250	32.38889	24.397046	88	0
2012901	7	31	0	8	712.200	0.000	135.849	31.06667	17.091908	66	5
2013002	1	29	0	1	420.163	0.000	24.000	31.48148	15.282897	70	5
2013002	2	81	3	5	1102.379	377.500	221.569	27.09091	13.628846	145	5
2013002	3	32	0	1	545.811	0.000	24.000	28.87500	16.163978	60	0
2013002	4	13	0	10	160.266	0.000	137.815	29.50000	23.114450	67	9
2013002	5	33	2	5	208.899	192.000	195.766	20.46154	10.974329	50	7
2013002	6	45	1	2	642.413	156.000	78.000	34.70270	22.067906	88	9
2013002	7	39	2	0	666.410	186.000	0.000	33.46154	19.704189	76	5
2013005	1	23	1	4	271.551	3.000	84.000	29.82353	13.130286	80	10
2013005	2	57	2	6	500.101	111.000	156.000	38.27273	17.718763	80	7
2013005	3	9	0	2	123.099	0.000	36.000	26.50000	21.407609	70	9
2013005	4	16	2	4	151.184	4.500	142.767	32.90000	17.922363	68	12
2013005	5	30	2	3	317.498	7.000	146.947	25.52381	10.424102	65	11
2013005	6	40	4	2	443.983	31.501	71.915	40.14815	25.673548	89	6
2013005	7	22	0	2	171.784	0.000	203.171	34.75000	16.625365	54	11

Species table

Number of fish (organized by family and species) caught by Gillnet or Traps for each survey, and in total across all surveys and gears.

• (new table for revision 2020) *

  fam_name	Sci_name	2012901_GN	2012901_TB	2013002_GN	2013002_TB	2013005_GN	2013005_TB	sum
  ACANTHURIDAE	Acanthurus nigricans	0	10	0	24	6	49	89
  ACANTHURIDAE	Acanthurus nigricauda	0	0	0	12	0	0	12
  SOLEIDAE	Achirus sp. *	0	0	0	0	1	0	1
  HOLOCENTRIDAE	Adioryx ruber	0	0	0	11	0	4	15
  HOLOCENTRIDAE	Adioryx spinifer	0	0	0	0	1	0	1
  SERRANIDAE	Aethaloperca rogaa	0	2	0	10	0	1	13
  ALBULIDAE	Albula vulpes	0	0	0	0	5	0	5
  CARANGIDAE	Alectis indicus	0	0	0	0	1	0	1
  CARANGIDAE	Alepes vari	0	0	0	0	4	0	4
  SPARIDAE	Argyrops filamentosus	0	0	0	7	0	0	7
  SPARIDAE	Argyrops sp.	0	20	0	0	0	0	20
  SPARIDAE	Argyrops spinifer	0	0	0	6	0	5	11
  ARIIDAE	Arius heudelotii	0	6	0	0	0	0	6
  ARIIDAE	Arius thalassinus	0	0	0	2	1	0	3
  SCOMBRIDAE	Auxis thazard	14	0	0	0	0	0	14
  BALISTIDAE	Balistapus undulatus	0	0	0	0	0	1	1
  BALISTIDAE	Balistoides viridescens	0	0	0	1	0	0	1
  BOTHIDAE	Bothus pantherinus	0	0	0	0	1	0	1
  CAESIONIDAE	Caesio caerulaurea	0	0	15	0	0	0	15
  CAESIONIDAE	Caesio suevicus	0	0	0	0	1	0	1
  CARANGIDAE	Carangoides armatus	0	0	0	0	4	0	4
  CARANGIDAE	Carangoides bajad	13	1	6	14	49	0	83
  CARANGIDAE	Carangoides ferdau	0	0	0	1	7	0	8
  CARANGIDAE	Carangoides fulvoguttatus	0	0	0	0	16	0	16
  CARANGIDAE	Carangoides sp.	0	0	0	0	1	0	1
  CARANGIDAE	Caranx (Caranx) melampygus	2	0	0	1	6	0	9
  CARANGIDAE	Caranx (Caranx) sexfasciatus	16	0	22	2	43	0	83
  CARANGIDAE	Caranx (Gnathanodon) speciosus	0	0	0	0	4	0	4
  CARANGIDAE	Caranx ignobilis	0	0	0	1	1	0	2
  CARANGIDAE	Caranx sp.	0	0	0	0	1	0	1
  CARCHARHINIDAE	Carcharhinus albimarginatus	2	0	0	0	0	0	2
  S H A R K S	Carcharhinus melanopterus	9	0	0	0	5	0	14
  S H A R K S	Carcharhinus wheeleri	0	0	1	0	1	0	2
  SERRANIDAE	Cephalopholis argus	0	1	0	0	0	0	1
  SERRANIDAE	Cephalopholis miniatus ?	0	0	0	0	1	0	1
  CHAETODONTIDAE	Chaetodon auriga	0	3	0	0	0	0	3
  CHAETODONTIDAE	Chaetodon semilarvatus	0	0	16	2	0	0	18
  CHANIDAE	Chanos chanos	0	0	0	0	1	0	1
  LABRIDAE	Cheilinus lunulatus	0	0	0	1	0	0	1
  LABRIDAE	Cheilinus quinquecintus	0	1	0	0	0	0	1
  CHIROCENTRIDAE	Chirocentrus dorab	15	0	17	0	60	0	92
  CARANGIDAE	Chorinemus lysan	0	0	0	6	0	0	6
  PLATYCEPHALIDAE	Cociella crocodila	0	0	0	0	1	0	1
  MUGILIDAE	Crenimugil crenilabis	0	0	1	0	0	0	1
  CARANGIDAE	Decapterus macarellus	0	0	1	0	0	0	1
  SCOMBRIDAE	Decapterus russelli	0	0	1	4	0	0	5
  HAEMULIDAE	Diagramma pictum	0	1	0	0	0	0	1
  DIODONTIDAE	Diodon hystrix	0	0	0	0	3	0	3
  ECHENEIDIDAE	Echeneis naucrates	0	0	0	3	2	0	5
  SCOMBRIDAE	Elagatis bipinnulata	0	0	6	0	0	0	6
  SERRANIDAE	Epinephelus chlorostigma	0	0	0	1	0	0	1
  SERRANIDAE	Epinephelus fuscoguttatus	0	7	0	7	1	2	17
  SERRANIDAE	Epinephelus sexfasciatus	0	0	0	1	0	1	2
  SERRANIDAE	Epinephelus summana	0	0	2	1	0	0	3
  SERRANIDAE	Epinephelus tauvina	1	11	4	4	3	2	25
  SCOMBRIDAE	Euthynnus affinis	0	0	4	0	3	0	7
  JUVENILES	FISHJUV	9	0	0	1	0	0	10
  GERREIDAE	Gerres oyena	0	0	1	0	6	0	7
  SCOMBRIDAE	Grammatorcynus bicarinatus	0	0	3	0	0	0	3
  SCOMBRIDAE	Grammatorcynus bilineatus	26	2	2	0	5	0	35
  LETHRINIDAE	Gymnocranius microdon	0	0	1	0	0	0	1
  LETHRINIDAE	Gymnocranius robinsoni	0	0	0	2	3	0	5
  SCOMBRIDAE	Gymnosarda unicolor	0	0	0	0	10	0	10
  MURAENIDAE	Gymnothorax flavimarginatus	0	3	0	0	0	0	3
  MURAENIDAE	Gymnothorax javanicus	0	28	0	11	0	1	40
  HAEMULIDAE	Hectromynicus pictus	0	0	0	4	0	0	4
  HEMIRAMPHIDAE	Hemiramphus far	0	0	0	0	1	0	1
  SCARIDAE	Hipposcarus harid	0	0	4	0	4	0	8
  SCOMBRIDAE	Katsuwonus pelamis	1	0	0	0	0	0	1
  KYPHOSIDAE	Kyphosus vaigiensis	0	0	0	0	10	0	10
  LETHRINIDAE	Lethrinus elongatus *	0	10	1	17	7	5	40
  LETHRINIDAE	Lethrinus harak	0	0	0	0	10	0	10
  LETHRINIDAE	Lethrinus lentjan	3	33	6	47	31	28	148
  LETHRINIDAE	Lethrinus mahsena	0	16	2	19	0	18	55
  LETHRINIDAE	Lethrinus mahsenoides *	0	4	0	0	6	0	10
  LETHRINIDAE	Lethrinus nebulosus	0	0	0	0	0	1	1
  LETHRINIDAE	Lethrinus ramak	0	0	1	1	0	1	3
  LETHRINIDAE	Lethrinus xanthochilus	0	0	0	2	0	1	3
  LUTJANIDAE	Lutjanus argentimaculatus	0	0	0	1	0	1	2
  LUTJANIDAE	Lutjanus bohar	0	32	10	69	3	8	122
  LUTJANIDAE	Lutjanus cf fluviflamma	0	0	0	0	2	0	2
  LUTJANIDAE	Lutjanus coccineus *	0	0	0	0	0	1	1
  LUTJANIDAE	Lutjanus ehrenbergii	6	0	11	0	7	0	24
  LUTJANIDAE	Lutjanus fulviflamma	0	0	0	0	2	0	2
  LUTJANIDAE	Lutjanus gibbus	0	38	0	51	1	28	118
  LUTJANIDAE	Lutjanus kasmira	0	4	0	3	0	2	9
  LUTJANIDAE	Lutjanus monostigma	2	4	0	3	0	1	10
  LUTJANIDAE	Lutjanus rivulatus	0	0	0	1	0	0	1
  LUTJANIDAE	Lutjanus sebae	0	0	0	1	0	0	1
  LUTJANIDAE	Lutjanus sp.	0	1	0	0	0	0	1
  LUTJANIDAE	Macolor niger	0	0	0	2	1	0	3
  LETHRINIDAE	Monotaxis grandoculis	0	0	1	0	0	0	1
  MULLIDAE	Mulloides flavolineatus	0	0	1	0	0	0	1
  MULLIDAE	Mulloides vanicolensis	0	0	0	0	1	0	1
  HOLOCENTRIDAE	Myripristis murdjan	0	0	3	0	4	0	7
  ACANTHURIDAE	Naso hexacanthus	0	0	18	18	37	0	73
  ACANTHURIDAE	Naso lituratus	0	0	0	0	1	0	1
  NO CATCH	NO CATCH	3	0	0	0	0	0	3
  LUTJANIDAE	Paracaesio sordius	0	2	0	0	0	0	2
  EPHIPPIDAE	Platax boersi	0	1	0	0	0	0	1
  EPHIPPIDAE	Platax orbicularis	0	2	0	2	1	2	7
  HAEMULIDAE	Plectorhinchus gaterinus	0	7	1	0	3	0	11
  POMADASYIDAE (HAEMULIDAE)	Plectorhinchus pictus	0	0	0	0	1	0	1
  HAEMULIDAE	Plectorhynchus pictus	0	0	0	0	1	0	1
  HAEMULIDAE	Plectorhynchus schotaf	0	0	0	0	1	0	1
  SERRANIDAE	Plectropomus pessuliferus	0	2	0	2	0	0	4
  PRIACANTHIDAE	Priacanthus hamrur	0	0	4	0	0	0	4
  LUTJANIDAE	Pristipomoides multidens	0	9	0	7	0	0	16
  BALISTIDAE	Pseudobalistes flavimarginatus	0	1	0	0	0	0	1
  SCOMBRIDAE	Rastrelliger kanagurta	0	0	6	0	21	0	27
  SCOMBRIDAE	Sarda orientalis	0	0	1	0	0	0	1
  HOLOCENTRIDAE	Sargocentron spiniferum	0	13	1	13	0	5	32
  SCARIDAE	Scarus ferrugineus	0	0	1	0	0	0	1
  SCARIDAE	Scarus frenatus	0	0	2	0	2	0	4
  SCARIDAE	Scarus ghobban	0	0	0	0	1	0	1
  SCOMBRIDAE	Scomber japonicus	0	0	0	1	0	0	1
  CARANGIDAE	Scomberoides lysan	15	20	60	18	45	0	158
  CARANGIDAE	Scomberoides tol	0	0	4	0	74	0	78
  SCOMBRIDAE	Scomberomorus commerson	35	0	0	0	0	0	35
  SCOMBRIDAE	Scomberomorus lineolatus	0	0	0	0	9	0	9
  SCOMBRIDAE	Scomberomorus tritor	2	0	0	0	0	0	2
  SIGANIDAE	Siganus argenteus	0	0	3	0	0	0	3
  SIGANIDAE	Siganus luridus	1	0	1	0	0	0	2
  SIGANIDAE	Siganus rivulatus	0	0	0	0	1	0	1
  SIGANIDAE	Siganus stellatus	0	0	1	0	0	2	3
  SPARIDAE	Sparus sp.	0	0	0	6	0	0	6
  SPHYRAENIDAE	Sphyraena forsteri	0	0	1	0	0	0	1
  SPHYRAENIDAE	Sphyraena jello	3	0	0	0	0	0	3
  SPHYRAENIDAE	Sphyraena putnamie	0	0	0	0	1	0	1
  SPHYRAENIDAE	Sphyraena qenie	0	0	6	0	4	0	10
  CARCHARHINIDAE	Sphyrna lewini	1	0	0	0	0	0	1
  R A Y S	Taeniura lymma	0	0	0	0	1	0	1
  SCOMBRIDAE	Thunnus albacares	11	0	0	0	0	0	11
  CARCHARHINIDAE	Triaenodon obesus	0	3	0	10	0	0	13
  BELONIDAE	Tylosurus choram	2	0	4	0	0	0	6
  MUGILIDAE	Valamugil engeli	0	0	0	0	1	0	1
  SERRANIDAE	Variola louti	0	0	0	2	0	0	2

---

Number of set traps per depth

Violin plot of depths at wich traps were set. I prefer this visualization of depth ranges of the traps, rather than a (boring) table.

Test differneces in depths in each area between years and between areas

## 
##  Shapiro-Wilk normality test
## 
## data:  c3$depth
## W = 0.90427, p-value < 2.2e-16

## 
##  Kruskal-Wallis rank sum test
## 
## data:  c3$depth and c3$id
## Kruskal-Wallis chi-squared = 36.877, df = 6, p-value = 1.861e-06

## Warning in posthoc.kruskal.nemenyi.test.default(c3$depth, c3$id, "Chisq"): Ties
## are present. Chi-sq was corrected for ties.

## 
##  Pairwise comparisons using Nemenyi-test with Chi-squared    
##                        approximation for independent samples 
## 
## data:  c3$depth and c3$id 
## 
##   1      2       3      4      5      6     
## 2 1.0000 -       -      -      -      -     
## 3 0.8599 0.6487  -      -      -      -     
## 4 0.9804 0.9477  1.0000 -      -      -     
## 5 0.0025 1.2e-05 0.3022 0.4513 -      -     
## 6 0.9943 0.9561  0.9883 0.9997 0.0075 -     
## 7 0.9983 0.9860  0.9835 0.9994 0.0117 1.0000
## 
## P value adjustment method: none

## [1] "2012901"
## 
##  Kruskal-Wallis rank sum test
## 
## data:  ci$depth and ci$id
## Kruskal-Wallis chi-squared = 29.371, df = 6, p-value = 5.174e-05

## Warning in posthoc.kruskal.nemenyi.test.default(ci$depth, ci$id, "Chisq"): Ties
## are present. Chi-sq was corrected for ties.

## 
##  Pairwise comparisons using Nemenyi-test with Chi-squared    
##                        approximation for independent samples 
## 
## data:  ci$depth and ci$id 
## 
##   1      2      3      4      5      6     
## 2 0.9991 -      -      -      -      -     
## 3 0.9060 0.4375 -      -      -      -     
## 4 0.6436 0.3884 0.9600 -      -      -     
## 5 0.1017 0.0011 0.7700 1.0000 -      -     
## 6 0.7221 0.1350 1.0000 0.9811 0.8655 -     
## 7 0.9290 0.4572 1.0000 0.9415 0.6429 0.9995
## 
## P value adjustment method: none 
## [1] "2013002"
## 
##  Kruskal-Wallis rank sum test
## 
## data:  ci$depth and ci$id
## Kruskal-Wallis chi-squared = 13.422, df = 6, p-value = 0.0368

## Warning in posthoc.kruskal.nemenyi.test.default(ci$depth, ci$id, "Chisq"): Ties
## are present. Chi-sq was corrected for ties.

## 
##  Pairwise comparisons using Nemenyi-test with Chi-squared    
##                        approximation for independent samples 
## 
## data:  ci$depth and ci$id 
## 
##   1     2     3     4     5     6    
## 2 0.999 -     -     -     -     -    
## 3 0.997 1.000 -     -     -     -    
## 4 1.000 1.000 1.000 -     -     -    
## 5 0.405 0.460 0.783 0.768 -     -    
## 6 0.999 0.893 0.918 0.999 0.078 -    
## 7 1.000 0.989 0.988 1.000 0.226 1.000
## 
## P value adjustment method: none 
## [1] "2013005"
## 
##  Kruskal-Wallis rank sum test
## 
## data:  ci$depth and ci$id
## Kruskal-Wallis chi-squared = 10.394, df = 6, p-value = 0.109

## Warning in posthoc.kruskal.nemenyi.test.default(ci$depth, ci$id, "Chisq"): Ties
## are present. Chi-sq was corrected for ties.

## 
##  Pairwise comparisons using Nemenyi-test with Chi-squared    
##                        approximation for independent samples 
## 
## data:  ci$depth and ci$id 
## 
##   1    2    3    4    5    6   
## 2 1.00 -    -    -    -    -   
## 3 1.00 0.98 -    -    -    -   
## 4 1.00 1.00 1.00 -    -    -   
## 5 0.64 0.13 0.99 0.85 -    -   
## 6 1.00 0.98 1.00 1.00 0.66 -   
## 7 1.00 1.00 1.00 1.00 0.68 1.00
## 
## P value adjustment method: none

Depth data were not normally distributed and hence the difference in depth could only be analyzed using non-parametric Kruskal - Wallis rank sum test.

Of the 63 survey - area comparisons, depth distributions were different between areas as follows: Nov.2012 survey: 1:5, 2:5, 5:6 and 5:7
May 2013 survey: 2:5

No significant differences for the Nov. 2013 survey

GAM model of trap depths

Fits a GAM model (using GAMLSS package) to the depth with area as explanatory variable and survey as a random factor. Use ‘chooseDIst’ function to test which distribution on the positive real line best fit the data.

## GAMLSS-RS iteration 1: Global Deviance = 5840.466 
## GAMLSS-RS iteration 2: Global Deviance = 5840.466

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## minimum GAIC(k= 2 ) family: exGAUS 
## minimum GAIC(k= 3.84 ) family: exGAUS 
## minimum GAIC(k= 6.51 ) family: exGAUS

##                 2     3.84     6.51
## EXP            NA       NA       NA
## GA             NA       NA       NA
## IG             NA       NA       NA
## LOGNO          NA       NA       NA
## LOGNO2         NA       NA       NA
## WEI            NA       NA       NA
## WEI2           NA       NA       NA
## WEI3           NA       NA       NA
## IGAMMA         NA       NA       NA
## PARETO2  6076.280 6091.000 6112.360
## PARETO2o 6028.005 6042.725 6064.085
## GP             NA       NA       NA
## BCCG           NA       NA       NA
## BCCGo          NA       NA       NA
## exGAUS   5676.647 5693.207 5717.237
## GG             NA       NA       NA
## GIG            NA       NA       NA
## LNO            NA       NA       NA
## BCTo           NA       NA       NA
## BCT            NA       NA       NA
## BCPEo          NA       NA       NA
## BCPE           NA       NA       NA
## GB2            NA       NA       NA

## GAIG with k= 6.51

##   exGAUS PARETO2o  PARETO2      EXP       GA       IG    LOGNO   LOGNO2 
## 5717.237 6064.085 6112.360       NA       NA       NA       NA       NA 
##      WEI     WEI2     WEI3   IGAMMA       GP     BCCG    BCCGo       GG 
##       NA       NA       NA       NA       NA       NA       NA       NA 
##      GIG      LNO     BCTo      BCT    BCPEo     BCPE      GB2 
##       NA       NA       NA       NA       NA       NA       NA

## GAIG with k= 6.51 
## GAMLSS-RS iteration 1: Global Deviance = 5702.051 
## GAMLSS-RS iteration 2: Global Deviance = 5682.573 
## GAMLSS-RS iteration 3: Global Deviance = 5669.27 
## GAMLSS-RS iteration 4: Global Deviance = 5662.855 
## GAMLSS-RS iteration 5: Global Deviance = 5660.216 
## GAMLSS-RS iteration 6: Global Deviance = 5659.202 
## GAMLSS-RS iteration 7: Global Deviance = 5658.841 
## GAMLSS-RS iteration 8: Global Deviance = 5658.714 
## GAMLSS-RS iteration 9: Global Deviance = 5658.675 
## GAMLSS-RS iteration 10: Global Deviance = 5658.656 
## GAMLSS-RS iteration 11: Global Deviance = 5658.65 
## GAMLSS-RS iteration 12: Global Deviance = 5658.648 
## GAMLSS-RS iteration 13: Global Deviance = 5658.647

## GAIG with k= 6.51 
## ******************************************************************
## Family:  c("exGAUS", "ex-Gaussian") 
## 
## Call:  gamlss(formula = depth ~ factor(id) + random(factor(survey)),  
##     family = names(getOrder(t1, 3)[1]), data = c3) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  identity
## Mu Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 15.15720    1.47773  10.257   <2e-16 ***
## factor(id)2  0.08425    1.66309   0.051   0.9596    
## factor(id)3 -3.55885    2.00296  -1.777   0.0761 .  
## factor(id)4 -4.87739    2.41340  -2.021   0.0437 *  
## factor(id)5 -3.79141    1.82704  -2.075   0.0384 *  
## factor(id)6 -3.68338    1.77533  -2.075   0.0384 *  
## factor(id)7 -2.10235    1.89522  -1.109   0.2677    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   1.7951     0.0917   19.57   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Nu link function:  log 
## Nu Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.92573    0.05178    56.5   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## NOTE: Additive smoothing terms exist in the formulas: 
##  i) Std. Error for smoothers are for the linear effect only. 
## ii) Std. Error for the linear terms maybe are not accurate. 
## ------------------------------------------------------------------
## No. of observations in the fit:  674 
## Degrees of Freedom for the fit:  9.000015
##       Residual Deg. of Freedom:  665 
##                       at cycle:  13 
##  
## Global Deviance:     5658.647 
##             AIC:     5676.647 
##             SBC:     5717.266 
## ******************************************************************

## 
## Family:  c("exGAUS", "ex-Gaussian") 
## Fitting method: RS() 
## 
## Call:  gamlss(formula = depth ~ factor(id) + random(factor(survey)),  
##     family = names(getOrder(t1, 3)[1]), data = c3) 
## 
## Mu Coefficients:
##             (Intercept)              factor(id)2              factor(id)3  
##                15.15720                  0.08425                 -3.55885  
##             factor(id)4              factor(id)5              factor(id)6  
##                -4.87739                 -3.79141                 -3.68338  
##             factor(id)7  random(factor(survey))1  
##                -2.10235                       NA  
## Sigma Coefficients:
## (Intercept)  
##       1.795  
## Nu Coefficients:
## (Intercept)  
##       2.926  
## 
##  Degrees of Freedom for the fit: 9 Residual Deg. of Freedom   665 
## Global Deviance:     5658.65 
##             AIC:     5676.65 
##             SBC:     5717.27

## [1] 5676.647

## GAIG with k= 6.51 
## GAIG with k= 6.51

## Single term deletions for
## mu
## 
## Model:
## depth ~ factor(id) + random(factor(survey))
##                                Df    AIC    LRT   Pr(Chi)    
## <none>                            5676.6                     
## factor(id)             6.0000e+00 5679.5 14.821   0.02169 *  
## random(factor(survey)) 1.5414e-05 5676.6  0.000 8.247e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## 
## Family:  c("exGAUS", "ex-Gaussian") 
## Fitting method: "nlminb" 
## 
## Call:  gamlssML(formula = depth, family = "exGAUS", data = c3) 
## 
## Mu Coefficients:
## [1]  12.55
## Sigma Coefficients:
## [1]  1.753
## Nu Coefficients:
## [1]  2.955
## 
##  Degrees of Freedom for the fit: 3 Residual Deg. of Freedom   671 
## Global Deviance:     5673.47 
##             AIC:     5679.47 
##             SBC:     5693.01

The GAMLSSS model picked ‘exGAUS’ as the best distribution to use in the model, which seems sensible from plotting the model over historgram of the depth.

Using the ‘exGAUS’ model areas 3, 4, 5 and 6 come out as significant factors, meaning the depth in these areas are significantly different from depths in other areas. The ‘drop1’ function showed that surveys as a random effect were a significant factor in the model.

Combination of GAMLSS model and Kruskal - Wallis

The GAMLSS model shows that there are signicant effects of area 3, 4, 5, and 6 on the depth distribution, and the K-W test shows significant differences for the Nov. 2012 survey between area 1:5, 2:5, 5:6, 5:7 and for the May 2013 survey for areas 2:4, which is in corrspondence with the GAM results.

---

Duration of trap sets

Mean, min & max hrs. of fishing

id	maxhrs	minhrs	meanhrs
1	46.417	11.500	20.48641
2	44.500	7.917	16.47669
3	24.250	8.167	16.86333
4	27.033	13.033	16.11076
5	40.200	11.667	17.79179
6	30.083	8.000	19.00436
7	31.866	10.417	19.47891

##        ss                  id             fhrs       
##  Min.   :2.013e+09   Min.   :1.000   Min.   : 7.917  
##  1st Qu.:2.013e+09   1st Qu.:2.000   1st Qu.:15.425  
##  Median :2.013e+09   Median :4.000   Median :16.150  
##  Mean   :2.013e+09   Mean   :3.909   Mean   :17.984  
##  3rd Qu.:2.013e+09   3rd Qu.:6.000   3rd Qu.:18.700  
##  Max.   :2.013e+09   Max.   :7.000   Max.   :46.417

Fishing hours (traps) by survey and area

Boxplot of median with mean values plotted as red circle. The average trap soak-time (fishing hours) across all surveys and areas was 17.984 hrs with a median of 16.15, but the November 2012 survey had more varied and higher soak times than the May 2013 and Nov. 2013.

Test differences in soak times (traps) K-W test

## 
##  Shapiro-Wilk normality test
## 
## data:  c3$Fhrs
## W = 0.67297, p-value < 2.2e-16

## 
##  Kruskal-Wallis rank sum test
## 
## data:  c3$Fhrs and c3$id
## Kruskal-Wallis chi-squared = 68.971, df = 6, p-value = 6.645e-13

## Warning in posthoc.kruskal.nemenyi.test.default(c3$Fhrs, c3$id, "Chisq"): Ties
## are present. Chi-sq was corrected for ties.

## 
##  Pairwise comparisons using Nemenyi-test with Chi-squared    
##                        approximation for independent samples 
## 
## data:  c3$Fhrs and c3$id 
## 
##   1       2       3       4       5       6      
## 2 0.89835 -       -       -       -       -      
## 3 1.00000 0.88439 -       -       -       -      
## 4 0.61831 0.95548 0.60026 -       -       -      
## 5 0.99997 0.65429 1.00000 0.42582 -       -      
## 6 0.14066 1.3e-05 0.20277 0.00167 0.19214 -      
## 7 0.01625 2.0e-07 0.02891 0.00012 0.02234 0.97141
## 
## P value adjustment method: none

## [1] "2012901"
## 
##  Kruskal-Wallis rank sum test
## 
## data:  ci$Fhrs and ci$id
## Kruskal-Wallis chi-squared = 60.806, df = 6, p-value = 3.087e-11

## Warning in posthoc.kruskal.nemenyi.test.default(ci$Fhrs, ci$id, "Chisq"): Ties
## are present. Chi-sq was corrected for ties.

## 
##  Pairwise comparisons using Nemenyi-test with Chi-squared    
##                        approximation for independent samples 
## 
## data:  ci$Fhrs and ci$id 
## 
##   1       2       3       4       5       6      
## 2 0.07376 -       -       -       -       -      
## 3 0.06785 0.99940 -       -       -       -      
## 4 0.18427 0.94372 0.98714 -       -       -      
## 5 0.96585 0.48177 0.39946 0.46876 -       -      
## 6 0.99285 0.00023 0.00085 0.04610 0.49836 -      
## 7 0.96104 8.6e-05 0.00033 0.02854 0.32747 0.99990
## 
## P value adjustment method: none 
## [1] "2013002"
## 
##  Kruskal-Wallis rank sum test
## 
## data:  ci$Fhrs and ci$id
## Kruskal-Wallis chi-squared = 42.845, df = 6, p-value = 1.252e-07

## Warning in posthoc.kruskal.nemenyi.test.default(ci$Fhrs, ci$id, "Chisq"): Ties
## are present. Chi-sq was corrected for ties.

## 
##  Pairwise comparisons using Nemenyi-test with Chi-squared    
##                        approximation for independent samples 
## 
## data:  ci$Fhrs and ci$id 
## 
##   1       2       3       4       5       6      
## 2 0.99993 -       -       -       -       -      
## 3 0.23422 0.14820 -       -       -       -      
## 4 0.99997 1.00000 0.74873 -       -       -      
## 5 0.37056 0.28089 0.99998 0.85136 -       -      
## 6 0.00475 0.00024 0.94518 0.20902 0.83861 -      
## 7 0.06190 0.01735 0.99966 0.51167 0.99399 0.99505
## 
## P value adjustment method: none 
## [1] "2013005"
## 
##  Kruskal-Wallis rank sum test
## 
## data:  ci$Fhrs and ci$id
## Kruskal-Wallis chi-squared = 27.538, df = 6, p-value = 0.0001148

## Warning in posthoc.kruskal.nemenyi.test.default(ci$Fhrs, ci$id, "Chisq"): Ties
## are present. Chi-sq was corrected for ties.

## 
##  Pairwise comparisons using Nemenyi-test with Chi-squared    
##                        approximation for independent samples 
## 
## data:  ci$Fhrs and ci$id 
## 
##   1     2     3     4     5     6    
## 2 0.375 -     -     -     -     -    
## 3 0.307 0.966 -     -     -     -    
## 4 0.962 0.995 0.870 -     -     -    
## 5 0.302 1.000 0.995 0.973 -     -    
## 6 0.979 0.830 0.627 1.000 0.719 -    
## 7 0.958 0.015 0.045 0.496 0.015 0.426
## 
## P value adjustment method: none

Of the 63 area- survey combinations 16 were significantly different (Kruskal Wallis rank sum test with Nemenyi post-hoc test, p<0.05). In Nov. 2012 area 1 was significantly different from 7, area 2 from 6 and7, area 3:7, area 4 vs 6 and 7, area 5 vs 6. In May 2013 area 2 was significantly different from 6 and 7, area 3 from 6 and 7, and area 4 from 6 and 7. In Nov 2013 area 1 was significantly different from area 6, while area 2 was significantly different from area 6 and 7.

GAMLSS model of soak times

## GAMLSS-RS iteration 1: Global Deviance = 3903.587 
## GAMLSS-RS iteration 2: Global Deviance = 3903.587

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## minimum GAIC(k= 2 ) family: BCPEo 
## minimum GAIC(k= 3.84 ) family: BCPEo 
## minimum GAIC(k= 6.51 ) family: BCPEo

## GAIG with k= 6.51

##    BCPEo     BCPE     BCTo      BCT      GB2       IG    BCCGo   exGAUS 
## 3257.667 3265.534 3304.446 3314.150 3407.192 3536.386 3559.944 3566.905 
##     BCCG       GG   IGAMMA      GIG    LOGNO      LNO       GA   LOGNO2 
## 3567.756 3570.111 3573.235 3579.743 3597.631 3597.631 3630.957 3686.328 
##      WEI     WEI3     WEI2      EXP PARETO2o       GP  PARETO2 
## 3788.868 3788.868 3803.669 5291.946 5331.031 5360.799 5367.490

## GAIG with k= 6.51 
## GAMLSS-RS iteration 1: Global Deviance = 3522.762 
## GAMLSS-RS iteration 2: Global Deviance = 3209.979 
## GAMLSS-RS iteration 3: Global Deviance = 3189.191 
## GAMLSS-RS iteration 4: Global Deviance = 3182.858 
## GAMLSS-RS iteration 5: Global Deviance = 3181.535 
## GAMLSS-RS iteration 6: Global Deviance = 3180.58 
## GAMLSS-RS iteration 7: Global Deviance = 3179.169 
## GAMLSS-RS iteration 8: Global Deviance = 3178.542 
## GAMLSS-RS iteration 9: Global Deviance = 3176.934 
## GAMLSS-RS iteration 10: Global Deviance = 3175.871 
## GAMLSS-RS iteration 11: Global Deviance = 3175.791 
## GAMLSS-RS iteration 12: Global Deviance = 3174.872 
## GAMLSS-RS iteration 13: Global Deviance = 3174.879 
## GAMLSS-RS iteration 14: Global Deviance = 3173.936 
## GAMLSS-RS iteration 15: Global Deviance = 3173.325 
## GAMLSS-RS iteration 16: Global Deviance = 3173.287 
## GAMLSS-RS iteration 17: Global Deviance = 3173.038 
## GAMLSS-RS iteration 18: Global Deviance = 3173.037

## GAMLSS-RS iteration 1: Global Deviance = 3903.547 
## GAMLSS-RS iteration 2: Global Deviance = 3903.547

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## minimum GAIC(k= 2 ) family: BCPEo 
## minimum GAIC(k= 3.84 ) family: BCPEo 
## minimum GAIC(k= 6.51 ) family: BCPEo

## GAIG with k= 6.51

##    BCPEo     BCPE     BCTo      BCT      GB2       IG    BCCGo   exGAUS 
## 3240.943 3259.051 3297.997 3307.693 3400.403 3530.032 3553.743 3560.860 
##     BCCG       GG   IGAMMA      GIG      LNO    LOGNO       GA   LOGNO2 
## 3561.571 3563.878 3566.942 3573.452 3591.307 3591.307 3624.615 3662.651 
##      WEI     WEI3     WEI2      EXP PARETO2o       GP  PARETO2 
## 3782.483 3782.483 3797.722 5285.641 5335.781 5354.508 5361.220

## GAIG with k= 6.51 
## GAMLSS-RS iteration 1: Global Deviance = 3523.817 
## GAMLSS-RS iteration 2: Global Deviance = 3211.688 
## GAMLSS-RS iteration 3: Global Deviance = 3188.282 
## GAMLSS-RS iteration 4: Global Deviance = 3183.423 
## GAMLSS-RS iteration 5: Global Deviance = 3178.531 
## GAMLSS-RS iteration 6: Global Deviance = 3176.571 
## GAMLSS-RS iteration 7: Global Deviance = 3174.21 
## GAMLSS-RS iteration 8: Global Deviance = 3171.155 
## GAMLSS-RS iteration 9: Global Deviance = 3169.13 
## GAMLSS-RS iteration 10: Global Deviance = 3168.107 
## GAMLSS-RS iteration 11: Global Deviance = 3165.407 
## GAMLSS-RS iteration 12: Global Deviance = 3164.784 
## GAMLSS-RS iteration 13: Global Deviance = 3164.143 
## GAMLSS-RS iteration 14: Global Deviance = 3163.747 
## GAMLSS-RS iteration 15: Global Deviance = 3163.593 
## GAMLSS-RS iteration 16: Global Deviance = 3163.362 
## GAMLSS-RS iteration 17: Global Deviance = 3162.806 
## GAMLSS-RS iteration 18: Global Deviance = 3162.682 
## GAMLSS-RS iteration 19: Global Deviance = 3162.855 
## GAMLSS-RS iteration 20: Global Deviance = 3162.823

## Warning in RS(): Algorithm RS has not yet converged

##           df      AIC
## fm2 12.00000 3186.823
## fm  12.99999 3199.037

## GAIG with k= 6.51

## Warning in summary.gamlss(fm2): summary: vcov has failed, option qr is used instead

## ******************************************************************
## Family:  c("BCPEo", "Box-Cox Power Exponential-orig.") 
## 
## Call:  
## gamlss(formula = Fhrs ~ factor(id) + factor(survey), family = names(getOrder(t1,  
##     3)[1]), data = c3) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  log
## Mu Coefficients:
##                         Estimate Std. Error  t value Pr(>|t|)    
## (Intercept)            2.9719240  0.0002393 12417.78   <2e-16 ***
## factor(id)2           -0.0146264  0.0002206   -66.29   <2e-16 ***
## factor(id)3           -0.0229227  0.0003061   -74.88   <2e-16 ***
## factor(id)4           -0.0344365  0.0002719  -126.67   <2e-16 ***
## factor(id)5           -0.0051845  0.0002663   -19.47   <2e-16 ***
## factor(id)6            0.1036512  0.0003240   319.92   <2e-16 ***
## factor(id)7            0.0323334  0.0002808   115.14   <2e-16 ***
## factor(survey)2013002 -0.1964314  0.0001841 -1067.17   <2e-16 ***
## factor(survey)2013005 -0.2068897  0.0003148  -657.24   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) -1.05359    0.06035  -17.46   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Nu link function:  identity 
## Nu Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  -2.9273     0.1482  -19.75   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Tau link function:  log 
## Tau Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) -0.90184    0.04022  -22.42   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## No. of observations in the fit:  674 
## Degrees of Freedom for the fit:  12
##       Residual Deg. of Freedom:  662 
##                       at cycle:  20 
##  
## Global Deviance:     3162.823 
##             AIC:     3186.823 
##             SBC:     3240.982 
## ******************************************************************

## 
## Family:  c("BCPEo", "Box-Cox Power Exponential-orig.") 
## Fitting method: RS() 
## 
## Call:  gamlss(formula = Fhrs ~ factor(id) + factor(survey),  
##     family = names(getOrder(t1, 3)[1]), data = c3) 
## 
## Mu Coefficients:
##           (Intercept)            factor(id)2            factor(id)3  
##              2.971924              -0.014626              -0.022923  
##           factor(id)4            factor(id)5            factor(id)6  
##             -0.034436              -0.005184               0.103651  
##           factor(id)7  factor(survey)2013002  factor(survey)2013005  
##              0.032333              -0.196431              -0.206890  
## Sigma Coefficients:
## (Intercept)  
##      -1.054  
## Nu Coefficients:
## (Intercept)  
##      -2.927  
## Tau Coefficients:
## (Intercept)  
##     -0.9018  
## 
##  Degrees of Freedom for the fit: 12 Residual Deg. of Freedom   662 
## Global Deviance:     3162.82 
##             AIC:     3186.82 
##             SBC:     3240.98

## GAIG with k= 6.51

## Warning in RS(): Algorithm RS has not yet converged

## GAIG with k= 6.51

## Warning in RS(): Algorithm RS has not yet converged

## Single term deletions for
## mu
## 
## Model:
## Fhrs ~ factor(id) + factor(survey)
##                Df    AIC     LRT   Pr(Chi)    
## <none>            3186.8                      
## factor(id)      6 3272.7  97.915 < 2.2e-16 ***
## factor(survey)  2 3406.3 223.455 < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## Warning in MLE(ll4, start = list(eta.mu = eta.mu, eta.sigma = eta.sigma, :
## possible convergence problem: optim gave code=1 false convergence (8)

## 
## Family:  c("BCPEo", "Box-Cox Power Exponential-orig.") 
## Fitting method: "nlminb" 
## 
## Call:  gamlssML(formula = Fhrs, family = "BCPEo", data = c3) 
## 
## Mu Coefficients:
## [1]  2.767
## Sigma Coefficients:
## [1]  -1.54
## Nu Coefficients:
## [1]  -1.495
## Tau Coefficients:
## [1]  -0.4741
## 
##  Degrees of Freedom for the fit: 4 Residual Deg. of Freedom   670 
## Global Deviance:     3448.48 
##             AIC:     3456.48 
##             SBC:     3474.53

GAMLSS found the ‘BCPEo’ to give the best fit to the ‘Fhrs’ data (see histogram plot).

Using the ‘BCPEo’ distribution a model of Fhrs by area (as factor) and survey as an ordinary factor showed that all surveys and areas had a significant effect on the soak time for traps, meaning that soak time varied siginificantly between the areas & surveys.

Duration and catch rates of gill-net sets

Fishing hours for traps was significantly affected by areas (except area 3 which had a very large variability), however, the number of gillnet sets was very low pr area and survey.

## GAMLSS-RS iteration 1: Global Deviance = 318.2773 
## GAMLSS-RS iteration 2: Global Deviance = 318.2772

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in RS(): Algorithm RS has not yet converged

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in RS(): Algorithm RS has not yet converged

## minimum GAIC(k= 2 ) family: BCPE 
## minimum GAIC(k= 3.84 ) family: BCPE 
## minimum GAIC(k= 4.16 ) family: BCPE

## GAIG with k= 4.16

##     BCPE    BCPEo       IG     BCCG    BCCGo       GG   IGAMMA   exGAUS 
## 170.3072 260.9338 338.0825 340.2585 340.3204 340.4890 342.4701 342.8582 
##    LOGNO      LNO      GB2       GA     WEI3      WEI     WEI2   LOGNO2 
## 343.0624 343.0624 343.8040 344.5514 355.4835 355.4847 355.5473 357.3183 
## PARETO2o      EXP       GP  PARETO2      GIG     BCTo      BCT 
## 512.5662 514.0901 524.0430 524.5871       NA       NA       NA

## GAIG with k= 4.16 
## GAMLSS-RS iteration 1: Global Deviance = 302.0237 
## GAMLSS-RS iteration 2: Global Deviance = 280.168 
## GAMLSS-RS iteration 3: Global Deviance = 257.7771 
## GAMLSS-RS iteration 4: Global Deviance = 227.273 
## GAMLSS-RS iteration 5: Global Deviance = 169.7975 
## GAMLSS-RS iteration 6: Global Deviance = 147.6317 
## GAMLSS-RS iteration 7: Global Deviance = 148.1086 
## GAMLSS-RS iteration 8: Global Deviance = 139.0477 
## GAMLSS-RS iteration 9: Global Deviance = 144.9091 
## GAMLSS-RS iteration 10: Global Deviance = 138.8345 
## GAMLSS-RS iteration 11: Global Deviance = 126.8551 
## GAMLSS-RS iteration 12: Global Deviance = 136.7751 
## GAMLSS-RS iteration 13: Global Deviance = 145.1821 
## GAMLSS-RS iteration 14: Global Deviance = 127.6621 
## GAMLSS-RS iteration 15: Global Deviance = 138.7439 
## GAMLSS-RS iteration 16: Global Deviance = 141.9076 
## GAMLSS-RS iteration 17: Global Deviance = 127.1116 
## GAMLSS-RS iteration 18: Global Deviance = 145.6108 
## GAMLSS-RS iteration 19: Global Deviance = 143.8334 
## GAMLSS-RS iteration 20: Global Deviance = 116.2272

## Warning in RS(): Algorithm RS has not yet converged

## GAMLSS-RS iteration 1: Global Deviance = 317.9713 
## GAMLSS-RS iteration 2: Global Deviance = 317.9713

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## minimum GAIC(k= 2 ) family: BCPE 
## minimum GAIC(k= 3.84 ) family: BCPE 
## minimum GAIC(k= 4.16 ) family: BCPE

## GAIG with k= 4.16

##     BCPE    BCPEo       IG     BCCG    BCCGo       GG   IGAMMA    LOGNO 
## 183.5930 219.4734 334.4460 337.0236 337.0703 337.4903 338.3114 339.4738 
##      LNO      GB2       GA   exGAUS      GIG   LOGNO2     WEI3      WEI 
## 339.4738 340.3919 340.9357 341.8466 342.4714 348.4250 351.7982 351.7982 
##     WEI2      EXP PARETO2o       GP  PARETO2     BCTo      BCT 
## 351.8761 510.7634 520.5020 520.8594 521.4184       NA       NA

## GAIG with k= 4.16 
## GAMLSS-RS iteration 1: Global Deviance = 302.5216 
## GAMLSS-RS iteration 2: Global Deviance = 284.742 
## GAMLSS-RS iteration 3: Global Deviance = 261.1607 
## GAMLSS-RS iteration 4: Global Deviance = 229.761 
## GAMLSS-RS iteration 5: Global Deviance = 186.3244 
## GAMLSS-RS iteration 6: Global Deviance = 171.3145 
## GAMLSS-RS iteration 7: Global Deviance = 145.1925 
## GAMLSS-RS iteration 8: Global Deviance = 134.7702 
## GAMLSS-RS iteration 9: Global Deviance = 133.1592 
## GAMLSS-RS iteration 10: Global Deviance = 134.4718 
## GAMLSS-RS iteration 11: Global Deviance = 144.9498 
## GAMLSS-RS iteration 12: Global Deviance = 142.6577 
## GAMLSS-RS iteration 13: Global Deviance = 136.4377 
## GAMLSS-RS iteration 14: Global Deviance = 125.8019 
## GAMLSS-RS iteration 15: Global Deviance = 116.6274 
## GAMLSS-RS iteration 16: Global Deviance = 142.7849 
## GAMLSS-RS iteration 17: Global Deviance = 123.6335 
## GAMLSS-RS iteration 18: Global Deviance = 120.8222 
## GAMLSS-RS iteration 19: Global Deviance = 114.3809 
## GAMLSS-RS iteration 20: Global Deviance = 133.673

## Warning in RS(): Algorithm RS has not yet converged

##     df      AIC
## fm  13 142.2272
## fm2 12 157.6730

## GAIG with k= 4.16 
## ******************************************************************
## Family:  c("BCPE", "Box-Cox Power Exponential") 
## 
## Call:  gamlss(formula = Fhrs ~ factor(id) + random(factor(survey)),  
##     family = names(getOrder(t1, 3)[1]), data = c33) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  identity
## Mu Coefficients:
##              Estimate Std. Error   t value Pr(>|t|)    
## (Intercept) 1.277e+01  3.907e-04 32690.955   <2e-16 ***
## factor(id)2 7.980e-01  5.837e-04  1367.302   <2e-16 ***
## factor(id)3 3.587e-05  6.669e-04     0.054    0.957    
## factor(id)4 4.980e-01  8.718e-04   571.242   <2e-16 ***
## factor(id)5 2.798e+00  8.712e-04  3211.871   <2e-16 ***
## factor(id)6 2.383e+00  6.607e-04  3606.873   <2e-16 ***
## factor(id)7 2.417e+00  1.065e-03  2268.835   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)    7.505      1.830   4.101 0.000148 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Nu link function:  identity 
## Nu Coefficients:
##             Estimate Std. Error t value Pr(>|t|)
## (Intercept)   -2.265      1.375  -1.648    0.106
## 
## ------------------------------------------------------------------
## Tau link function:  log 
## Tau Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  -2.9424     0.1471     -20   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## NOTE: Additive smoothing terms exist in the formulas: 
##  i) Std. Error for smoothers are for the linear effect only. 
## ii) Std. Error for the linear terms maybe are not accurate. 
## ------------------------------------------------------------------
## No. of observations in the fit:  64 
## Degrees of Freedom for the fit:  13
##       Residual Deg. of Freedom:  51 
##                       at cycle:  20 
##  
## Global Deviance:     116.2272 
##             AIC:     142.2272 
##             SBC:     170.2927 
## ******************************************************************

## GAIG with k= 4.16 
## GAIG with k= 4.16

## Warning in RS(): Algorithm RS has not yet converged

## Single term deletions for
## mu
## 
## Model:
## Fhrs ~ factor(id) + random(factor(survey))
##                            Df    AIC     LRT   Pr(Chi)    
## <none>                        142.23                      
## factor(id)             6.3179 304.18 174.586 < 2.2e-16 ***
## random(factor(survey)) 3.0000 197.11  60.882  3.81e-13 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## Warning in MLE(ll4, start = list(eta.mu = eta.mu, eta.sigma = eta.sigma, :
## possible convergence problem: optim gave code=1 false convergence (8)

## 
## Family:  c("BCPE", "Box-Cox Power Exponential") 
## Fitting method: "nlminb" 
## 
## Call:  gamlssML(formula = Fhrs, family = "BCPE", data = c3) 
## 
## Mu Coefficients:
## [1]  15.88
## Sigma Coefficients:
## [1]  -1.475
## Nu Coefficients:
## [1]  -0.8181
## Tau Coefficients:
## [1]  -0.5107
## 
##  Degrees of Freedom for the fit: 4 Residual Deg. of Freedom   670 
## Global Deviance:     3451.03 
##             AIC:     3459.03 
##             SBC:     3477.08

Both area and survey were found to be significan variables for explaining fishing hours of gillnets. when the smoothing function was dropped (‘drop’ function).

---

Analysis of catches by traits and other factors

Catch rates pr area & survey

Must aggregate data per station

Statistical analysis CPUE catches

Testing for normality of CPUEw of trap catches

Shapiro test for normality and Q-Q plots shows that data is non-normal and zero-inflated.

## 
##  Shapiro-Wilk normality test
## 
## data:  cpue.tr.st$CPUEw
## W = 0.55028, p-value < 2.2e-16

## 
## Attaching package: 'qqplotr'

## The following objects are masked from 'package:ggplot2':
## 
##     stat_qq_line, StatQqLine

### Test for difference in CPUEw of traps between areas (non-parametric test) Both CPUEw is not significantly different between the areas,

Significant difference for CPUEW between Nov 2013 survey and the Nov 2012 and May 2013 surveys, but not between the Nov12 and May13 surveys

Zero-inflated GAM model

This method is more appropriate than the Kruskal-Wallis approach above as it evaluates the effects of all factors at the same time, instead of evaluating the effects of area separate from survey.

First, developed a model for CPUE (weight and numbers) including depth, area, survey and gear (only traps and gillnets), which performed better than a simpler model excluding gear.

Also tested models where survey was included as a random factor (using two different approaches: ´random´and ´re´), although this excludes evaluation of the potential significant effects of the surveys on the results.

Tested models for both CPUE by weight (CPUEw) and CPUE by numbers (CPUEn).

Model performance were compared based on AIC.

Model specified:
CPUEw models:
* mod1<-gamlss(CPUEw~depth+factor(id)+factor(survey)+factor(gear),family=ZAGA,data=cpue.st)
* mod2<-gamlss(CPUEw~depth+factor(id)+factor(gear)+random(factor(survey)),family=ZAGA, data=cpue.st) * mod3<-gamlss(CPUEw_{depth+factor(id)+factor(gear)+re(random=}1|survey),family=ZAGA, data=cpue.st) * mod4 <-gamlss(CPUEw_{depth+factor(id)+factor(gear)+random(factor(survey)),nu.fo=}depth+factor(id)+factor(gear)+random(factor(survey)), family=ZAGA, data=cpue.st)
* mod5 <-gamlss(CPUEw~depth+factor(id)+factor(gear)+random(factor(survey)), family=ZAIG, data=cpue.st)
* mod6 <-gamlss(CPUEw_{depth+factor(id)+factor(gear)+factor(survey),nu.fo=}depth+factor(id)+factor(gear)+factor(survey), family=ZAGA, data=cpue.st)

Weight with Fhrs as dependent models: * modA1<-gamlss(weight~depth+Fhrs+factor(id)+factor(survey)+factor(gear),family=ZAGA, data=cpue.st)
* modA2 <- gamlss(weight~depth+Fhrs+factor(id)+random(factor(survey))+factor(gear),family=ZAGA, data=cpue.st)

Numbers with Fhrs as dependent model: * modA4 <- gamlss(number~depth+Fhrs+factor(id)+factor(survey)+factor(gear),family=ZAGA, data=cpue.st)

CPUEn models: * modA3<-gamlss(CPUEn~depth+factor(id)+factor(survey)+factor(gear),family=ZAGA, data=cpue.st)
* modA5<-gamlss(CPUEn~depth+factor(id)+random(factor(survey))+factor(gear),family=ZAGA, data=cpue.st)
* modA6 <- gamlss(CPUEn_{depth+factor(id)+random(factor(survey))+factor(gear),nu.fo=}depth+factor(id)+factor(gear)+random(factor(survey)),family=ZAGA, data=cpue.st)
* modA7 <- gamlss(CPUEn_{depth+factor(id)+factor(survey)+factor(gear),nu.fo=}depth+factor(id)+factor(gear)+factor(survey),family=ZAGA, data=cpue.st)

## GAMLSS-RS iteration 1: Global Deviance = 882.3705 
## GAMLSS-RS iteration 2: Global Deviance = 882.3705

## GAMLSS-RS iteration 1: Global Deviance = 883.5556 
## GAMLSS-RS iteration 2: Global Deviance = 883.5553

## GAMLSS-RS iteration 1: Global Deviance = 883.5553 
## GAMLSS-RS iteration 2: Global Deviance = 883.555

## GAMLSS-RS iteration 1: Global Deviance = 824.5923 
## GAMLSS-RS iteration 2: Global Deviance = 824.5951 
## GAMLSS-RS iteration 3: Global Deviance = 824.595

## GAMLSS-RS iteration 1: Global Deviance = 891.7485 
## GAMLSS-RS iteration 2: Global Deviance = 891.7485

## GAMLSS-RS iteration 1: Global Deviance = 822.9233 
## GAMLSS-RS iteration 2: Global Deviance = 822.9233

##            df      AIC
## mod4 23.29925 871.1935
## mod1 13.00000 908.3705
## mod3 12.73369 909.0224
## mod2 12.73357 909.0224
## mod5 13.72109 919.1906

## GAMLSS-RS iteration 1: Global Deviance = 2954.59 
## GAMLSS-RS iteration 2: Global Deviance = 2954.59

## GAMLSS-RS iteration 1: Global Deviance = 2955.38 
## GAMLSS-RS iteration 2: Global Deviance = 2955.379

## GAMLSS-RS iteration 1: Global Deviance = 622.7372 
## GAMLSS-RS iteration 2: Global Deviance = 622.7371

## GAMLSS-RS iteration 1: Global Deviance = 2653.48 
## GAMLSS-RS iteration 2: Global Deviance = 2653.48

## GAMLSS-RS iteration 1: Global Deviance = 624.8117 
## GAMLSS-RS iteration 2: Global Deviance = 624.8112

## GAMLSS-RS iteration 1: Global Deviance = 568.6383 
## GAMLSS-RS iteration 2: Global Deviance = 568.6426 
## GAMLSS-RS iteration 3: Global Deviance = 568.6427

## GAMLSS-RS iteration 1: Global Deviance = 566.0111 
## GAMLSS-RS iteration 2: Global Deviance = 566.0111

##             df       AIC
## modA7 23.00000  612.0111
## modA6 22.79631  614.2354
## modA3 13.00000  648.7371
## modA5 12.29026  649.3918
## mod6  23.00000  868.9233
## mod4  23.29925  871.1935
## mod1  13.00000  908.3705
## mod3  12.73369  909.0224
## mod2  12.73357  909.0224
## modA4 14.00000 2681.4797
## modA1 14.00000 2982.5900
## modA2 14.20733 2983.7937

## ******************************************************************
## Family:  c("ZAGA", "Zero Adjusted GA") 
## 
## Call:  gamlss(formula = CPUEw ~ depth + factor(id) + factor(gear) +  
##     factor(survey), nu.formula = ~depth + factor(id) +  
##     factor(gear) + factor(survey), family = ZAGA, data = cpue.st) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  log
## Mu Coefficients:
##                        Estimate Std. Error t value Pr(>|t|)    
## (Intercept)            0.353996   0.260322   1.360   0.1743    
## depth                 -0.002375   0.002834  -0.838   0.4022    
## factor(id)2            0.201090   0.193676   1.038   0.2995    
## factor(id)3           -0.010450   0.243952  -0.043   0.9658    
## factor(id)4           -0.352511   0.268869  -1.311   0.1903    
## factor(id)5           -0.285698   0.229247  -1.246   0.2131    
## factor(id)6           -0.175741   0.204906  -0.858   0.3914    
## factor(id)7           -0.213201   0.221633  -0.962   0.3364    
## factor(gear)TB        -1.554940   0.182336  -8.528   <2e-16 ***
## factor(survey)2013002  0.017219   0.127035   0.136   0.8922    
## factor(survey)2013005 -0.286449   0.162254  -1.765   0.0779 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)
## (Intercept)  0.02653    0.03243   0.818    0.414
## 
## ------------------------------------------------------------------
## Nu link function:  logit 
## Nu Coefficients:
##                        Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -1.891862   0.422176  -4.481 8.65e-06 ***
## depth                 -0.003244   0.004102  -0.791 0.429338    
## factor(id)2            0.186608   0.272054   0.686 0.492985    
## factor(id)3            0.278255   0.336710   0.826 0.408858    
## factor(id)4            0.376375   0.396877   0.948 0.343277    
## factor(id)5            0.459413   0.314411   1.461 0.144407    
## factor(id)6           -0.418105   0.299124  -1.398 0.162621    
## factor(id)7            0.321032   0.310490   1.034 0.301510    
## factor(gear)TB         1.749920   0.337799   5.180 2.89e-07 ***
## factor(survey)2013002  0.033728   0.184268   0.183 0.854821    
## factor(survey)2013005  0.775948   0.203050   3.821 0.000144 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## No. of observations in the fit:  734 
## Degrees of Freedom for the fit:  23
##       Residual Deg. of Freedom:  711 
##                       at cycle:  2 
##  
## Global Deviance:     822.9233 
##             AIC:     868.9233 
##             SBC:     974.689 
## ******************************************************************

## Single term deletions for
## mu
## 
## Model:
## CPUEw ~ depth + factor(id) + factor(gear) + factor(survey)
##                Df    AIC    LRT Pr(Chi)    
## <none>            868.92                   
## depth           1 867.60  0.677 0.41074    
## factor(id)      6 869.18 12.254 0.05654 .  
## factor(gear)    1 947.13 80.202 < 2e-16 ***
## factor(survey)  2 869.07  4.151 0.12551    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## ******************************************************************
##   Summary of the Randomised Quantile Residuals
##                            mean   =  0.01345799 
##                        variance   =  0.9924864 
##                coef. of skewness  =  0.01011289 
##                coef. of kurtosis  =  3.508819 
## Filliben correlation coefficient  =  0.99746 
## ******************************************************************

## ******************************************************************
## Family:  c("ZAGA", "Zero Adjusted GA") 
## 
## Call:  gamlss(formula = CPUEn ~ depth + factor(id) + factor(survey) +  
##     factor(gear), nu.formula = ~depth + factor(id) +  
##     factor(gear) + factor(survey), family = ZAGA, data = cpue.st) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  log
## Mu Coefficients:
##                        Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -0.111560   0.212432  -0.525   0.5996    
## depth                 -0.005006   0.002471  -2.026   0.0432 *  
## factor(id)2            0.343525   0.167665   2.049   0.0408 *  
## factor(id)3            0.076865   0.213978   0.359   0.7195    
## factor(id)4            0.190577   0.234250   0.814   0.4162    
## factor(id)5            0.467716   0.198977   2.351   0.0190 *  
## factor(id)6           -0.216463   0.178400  -1.213   0.2254    
## factor(id)7            0.246084   0.191877   1.283   0.2001    
## factor(survey)2013002  0.205185   0.109024   1.882   0.0602 .  
## factor(survey)2013005  0.284145   0.132690   2.141   0.0326 *  
## factor(gear)TB        -1.840150   0.151532 -12.144   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) -0.14492    0.03367  -4.304 1.92e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Nu link function:  logit 
## Nu Coefficients:
##                         Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -1.8154895  0.4213267  -4.309 1.87e-05 ***
## depth                 -0.0007469  0.0040111  -0.186 0.852337    
## factor(id)2            0.0469735  0.2716795   0.173 0.862779    
## factor(id)3            0.1816591  0.3366140   0.540 0.589596    
## factor(id)4            0.2177014  0.3955499   0.550 0.582234    
## factor(id)5            0.3332084  0.3138607   1.062 0.288758    
## factor(id)6           -0.5369061  0.2982262  -1.800 0.072232 .  
## factor(id)7            0.1664222  0.3100249   0.537 0.591572    
## factor(gear)TB         1.7318121  0.3362045   5.151 3.36e-07 ***
## factor(survey)2013002  0.0896906  0.1839197   0.488 0.625940    
## factor(survey)2013005  0.7543930  0.2027639   3.721 0.000214 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## No. of observations in the fit:  734 
## Degrees of Freedom for the fit:  23
##       Residual Deg. of Freedom:  711 
##                       at cycle:  2 
##  
## Global Deviance:     566.0111 
##             AIC:     612.0111 
##             SBC:     717.7768 
## ******************************************************************

## ******************************************************************
##   Summary of the Randomised Quantile Residuals
##                            mean   =  -0.01431757 
##                        variance   =  1.031279 
##                coef. of skewness  =  0.3962359 
##                coef. of kurtosis  =  4.647862 
## Filliben correlation coefficient  =  0.9888994 
## ******************************************************************

## Single term deletions for
## mu
## 
## Model:
## CPUEn ~ depth + factor(id) + factor(survey) + factor(gear)
##                Df    AIC     LRT   Pr(Chi)    
## <none>            612.01                      
## depth           1 613.90   3.891 0.0485368 *  
## factor(id)      6 623.53  23.520 0.0006399 ***
## factor(survey)  2 613.41   5.404 0.0670865 .  
## factor(gear)    1 768.46 158.453 < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## 
## Family:  c("ZAGA", "Zero Adjusted GA") 
## Fitting method: RS() 
## 
## Call:  gamlss(formula = CPUEw ~ depth + factor(id) + factor(gear) +  
##     factor(survey), nu.formula = ~depth + factor(id) +  
##     factor(gear) + factor(survey), family = ZAGA, data = cpue.st) 
## 
## Mu Coefficients:
##           (Intercept)                  depth            factor(id)2  
##              0.353996              -0.002375               0.201090  
##           factor(id)3            factor(id)4            factor(id)5  
##             -0.010450              -0.352511              -0.285698  
##           factor(id)6            factor(id)7         factor(gear)TB  
##             -0.175741              -0.213201              -1.554940  
## factor(survey)2013002  factor(survey)2013005  
##              0.017219              -0.286449  
## Sigma Coefficients:
## (Intercept)  
##     0.02653  
## Nu Coefficients:
##           (Intercept)                  depth            factor(id)2  
##             -1.891862              -0.003244               0.186608  
##           factor(id)3            factor(id)4            factor(id)5  
##              0.278255               0.376375               0.459413  
##           factor(id)6            factor(id)7         factor(gear)TB  
##             -0.418105               0.321032               1.749920  
## factor(survey)2013002  factor(survey)2013005  
##              0.033728               0.775948  
## 
##  Degrees of Freedom for the fit: 23 Residual Deg. of Freedom   711 
## Global Deviance:     822.923 
##             AIC:     868.923 
##             SBC:     974.689

## 
## Family:  c("ZAGA", "Zero Adjusted GA") 
## Fitting method: RS() 
## 
## Call:  gamlss(formula = CPUEn ~ depth + factor(id) + factor(survey) +  
##     factor(gear), nu.formula = ~depth + factor(id) +  
##     factor(gear) + factor(survey), family = ZAGA, data = cpue.st) 
## 
## Mu Coefficients:
##           (Intercept)                  depth            factor(id)2  
##             -0.111560              -0.005006               0.343525  
##           factor(id)3            factor(id)4            factor(id)5  
##              0.076865               0.190577               0.467716  
##           factor(id)6            factor(id)7  factor(survey)2013002  
##             -0.216463               0.246084               0.205185  
## factor(survey)2013005         factor(gear)TB  
##              0.284145              -1.840150  
## Sigma Coefficients:
## (Intercept)  
##     -0.1449  
## Nu Coefficients:
##           (Intercept)                  depth            factor(id)2  
##            -1.8154895             -0.0007469              0.0469735  
##           factor(id)3            factor(id)4            factor(id)5  
##             0.1816591              0.2177014              0.3332084  
##           factor(id)6            factor(id)7         factor(gear)TB  
##            -0.5369061              0.1664222              1.7318121  
## factor(survey)2013002  factor(survey)2013005  
##             0.0896906              0.7543930  
## 
##  Degrees of Freedom for the fit: 23 Residual Deg. of Freedom   711 
## Global Deviance:     566.011 
##             AIC:     612.011 
##             SBC:     717.777

	df	AIC
modA7	23.00000	612.0111
modA6	22.79631	614.2354
modA3	13.00000	648.7371
modA5	12.29026	649.3918
mod6	23.00000	868.9233
mod4	23.29925	871.1935
mod1	13.00000	908.3705
mod3	12.73369	909.0224
mod2	12.73357	909.0224
modA4	14.00000	2681.4797
modA1	14.00000	2982.5900
modA2	14.20733	2983.7937

Choice of models

• CPUEn models had lower AIC than CPUEw
• models with nu function specified performed better (lower AIC) than models without (for both CPUEw and CPUEn)
  
• model with survey as a random factor performed worse (higher AIC) than models with survey as an ordinary factor (for both CPUEw and CPUEn, as well as models with nu function defined)

Model for weight of catch (CPUEw)

• The best fitting (lowest AIC) model was ‘mod6’ where survey was an ordinary factor and the nu-function was defined.

• Area and gear were the only two variables found to be significant for the model when the smoothing function was dropped (‘drop1’ function)

• For mod6 ‘gear’ and ‘survey’ (Nov 2013) were significant factors. (intercept and depth were not)

Model for number of catch (CPUEn)

• The best fitting (lowest AIC) model was ‘modA7’ where survey was an ordinary factor and the nu-function specified.

• For modA7 the variables depth, area 1, area 2, area 5, surveys (all) and gear were significant factors.

• All variable (depth, area, survey and gear) were found to be significant variables in the model when the smoothing function was dropped (‘drop1’ function)

GAM model CPUE - traps only

Statistical analysis of plot of trap catch rates

## GAMLSS-RS iteration 1: Global Deviance = 680.1101 
## GAMLSS-RS iteration 2: Global Deviance = 680.1094

## GAMLSS-RS iteration 1: Global Deviance = 679.8366 
## GAMLSS-RS iteration 2: Global Deviance = 679.8366

## GAMLSS-RS iteration 1: Global Deviance = 395.6543 
## GAMLSS-RS iteration 2: Global Deviance = 395.6522 
## GAMLSS-RS iteration 3: Global Deviance = 395.6523

## GAMLSS-RS iteration 1: Global Deviance = 394.265 
## GAMLSS-RS iteration 2: Global Deviance = 394.2647

##             df      AIC
## mod55 12.00000 418.2647
## mod44 11.77455 419.2014
## mod33 12.00000 703.8366
## mod22 12.66004 705.4294

## ******************************************************************
## Family:  c("ZAGA", "Zero Adjusted GA") 
## 
## Call:  gamlss(formula = CPUEn ~ depth + factor(id) + factor(survey),  
##     family = ZAGA, data = c22) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  log
## Mu Coefficients:
##                         Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -2.1132812  0.1940001 -10.893  < 2e-16 ***
## depth                 -0.0008637  0.0026047  -0.332  0.74030    
## factor(id)2            0.2826635  0.1679995   1.683  0.09294 .  
## factor(id)3            0.1774132  0.2149544   0.825  0.40947    
## factor(id)4            0.0003821  0.2591314   0.001  0.99882    
## factor(id)5            0.5122892  0.2040252   2.511  0.01228 *  
## factor(id)6           -0.2127678  0.1785497  -1.192  0.23383    
## factor(id)7            0.1951116  0.1968309   0.991  0.32192    
## factor(survey)2013002  0.2953217  0.1088671   2.713  0.00685 ** 
## factor(survey)2013005  0.2910664  0.1377350   2.113  0.03496 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) -0.20337    0.03668  -5.544 4.29e-08 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Nu link function:  logit 
## Nu Coefficients:
##             Estimate Std. Error t value Pr(>|t|)  
## (Intercept)  0.16154    0.07752   2.084   0.0376 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## No. of observations in the fit:  670 
## Degrees of Freedom for the fit:  12
##       Residual Deg. of Freedom:  658 
##                       at cycle:  2 
##  
## Global Deviance:     394.2647 
##             AIC:     418.2647 
##             SBC:     472.3521 
## ******************************************************************

## Single term deletions for
## mu
## 
## Model:
## CPUEn ~ depth + factor(id) + factor(survey)
##                Df    AIC     LRT  Pr(Chi)   
## <none>            418.26                    
## depth           1 416.38  0.1118 0.738110   
## factor(id)      6 428.57 22.3066 0.001065 **
## factor(survey)  2 422.18  7.9179 0.019084 * 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## 
## Family:  c("ZAGA", "Zero Adjusted GA") 
## Fitting method: RS() 
## 
## Call:  gamlss(formula = CPUEn ~ depth + factor(id) + factor(survey),  
##     family = ZAGA, data = c22) 
## 
## Mu Coefficients:
##           (Intercept)                  depth            factor(id)2  
##            -2.1132812             -0.0008637              0.2826635  
##           factor(id)3            factor(id)4            factor(id)5  
##             0.1774132              0.0003821              0.5122892  
##           factor(id)6            factor(id)7  factor(survey)2013002  
##            -0.2127678              0.1951116              0.2953217  
## factor(survey)2013005  
##             0.2910664  
## Sigma Coefficients:
## (Intercept)  
##     -0.2034  
## Nu Coefficients:
## (Intercept)  
##      0.1615  
## 
##  Degrees of Freedom for the fit: 12 Residual Deg. of Freedom   658 
## Global Deviance:     394.265 
##             AIC:     418.265 
##             SBC:     472.352

The CPUEn models had better fit than the CPUEw models.

For the model (mod55) of CPUEn with survey as an ordniary factor both area and survey were significant variable (when excluding the smoothing function), with all surveys and areas 1, 2 and 5 being significant factors in the model

---

Bottom depth traps and gillnets

## 
##  Welch Two Sample t-test
## 
## data:  subset(cpue.st, gear == "TB")$depth and (subset(cpue.st, gear == "GN")$depth)
## t = 3.5683, df = 67.859, p-value = 0.0006648
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##   6.034582 21.347788
## sample estimates:
## mean of x mean of y 
##  31.73806  18.04688

Depth was, however, not a significant factor in any of the four models developed for the trap CPUE, by weight or numbers. The depth distribution of bottom depth at the trap and gillnet stations (Gillnet fishing depth was at the surface, while trap fishing depth = bottom depth) were significantly different (t-test, p<0.05, df=67.9 - see also frequncy plot above). Therefore, the models including both gear types led to a wider span and a more diverse distribution of the depths than for the traps-only data set.

---

Combine duration and depth plots of traps in one panels

Use ‘patchwork’ package to arrange plots

## Saving 7 x 5 in image
## Saving 7 x 5 in image

---

CPUE by gear type - by family and trophic group

Only for gillnets and traps and stations with catches (excluding no-catch)

## Warning: Ignoring unknown parameters: binwidt

## Warning: Ignoring unknown parameters: binwidt

Multivariate analysis of catch compososition

PCA analysis was carried out to exlore the variability in the catch data.

PCA was carried out on average catches (CPUE) per survey and area (needed to create a M*N matrix - cannot have uneven number of observations per group).

##                     PC1         PC2         PC3        PC4           PC5 gear
## GN_2012901_2 -0.1290066 -0.08358818  0.03147990 0.08337330 -1.077753e-04   GN
## GN_2012901_3 -0.1312917 -0.12467455  0.02572704 0.08334503  4.303331e-05   GN
## GN_2012901_5 -0.1310277 -0.13175803  0.03264853 0.08441385  1.259408e-05   GN
## GN_2012901_6 -0.1364297  0.01319564 -0.10899020 0.06254201  6.354907e-04   GN
## GN_2012901_7 -0.1298589 -0.16312191  0.06329514 0.08914630 -1.221831e-04   GN
## GN_2013002_1 -0.1277763  0.06995394 -0.16579955 0.05381124  8.921758e-04   GN
##               survey area
## GN_2012901_2 2012901    2
## GN_2012901_3 2012901    3
## GN_2012901_5 2012901    5
## GN_2012901_6 2012901    6
## GN_2012901_7 2012901    7
## GN_2013002_1 2013002    1

## 
## Attaching package: 'gridExtra'

## The following object is masked from 'package:dplyr':
## 
##     combine

##                   PC1         PC2          PC3          PC4          PC5
## Carni.   0.0264913831 -0.71085582  0.694599293  0.107259939 -0.003054698
## Herb.    0.0892255565  0.70172234  0.701988398  0.082398331 -0.006881644
## Plankt.  0.9956579132 -0.04394277 -0.081589489 -0.008747910  0.000991115
## Invert.  0.0015000416 -0.01823351  0.134225669 -0.990728833 -0.010253712
## Coral.  -0.0002765113  0.00251435  0.008410486 -0.009256049  0.999918592
##         Trophic_Group
## Carni.         Carni.
## Herb.           Herb.
## Plankt.       Plankt.
## Invert.       Invert.
## Coral.         Coral.

Combine gearplot and PCA

PCA of catch weight and numbers (CPUEw and CPUEn) by trophic group gave similar results and showed how trap and gillnet catches differed. The first two principal components explained 78.7% of the variation. Catches of Planktivores drove the variability along PC1 with gillnet catches in area 2 during the Nov. 2013 survey having high catches of planktivores compared to the other data points. PC2 was driven by the catches of herbivores which had high catches in the traps, and carnivores with higher catches in the gillnets. Catches of coralivores and invertivores did not contribute to the variability in catches (but these were caught in low quantities in just a few surveys and areas).

Conclusion: Trap catches were dominated by herbivores, gillnet catches by carnivores, while gillnet catches in area 2 during the Nov. 2013 survey had very high catches of planktivores compared to other gears, surveys and areas.

However, this is still descriptive. Need to evaluate the differences statistically. Will use GAM model.
(This also allows to use data on station level, not just average catches per area/survey needed to construct a matrix necessary for PCA)

## GAMLSS-RS iteration 1: Global Deviance = -95.5115 
## GAMLSS-RS iteration 2: Global Deviance = -95.5115

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## minimum GAIC(k= 2 ) family: GIG 
## minimum GAIC(k= 3.84 ) family: GIG 
## minimum GAIC(k= 6.19 ) family: GIG

## GAIG with k= 6.19

##        GIG        LNO      LOGNO         GG       BCTo        GB2         IG 
##  -833.2172  -830.7056  -830.7056  -827.4265  -821.0149  -809.4175  -803.0082 
##     IGAMMA    PARETO2         GP   PARETO2o       WEI3        WEI       WEI2 
##  -801.7104  -785.8249  -785.3709  -784.5883  -763.7464  -763.7464  -763.7462 
##         GA        EXP      BCPEo       BCPE      BCCGo     LOGNO2       BCCG 
##  -753.2317  -752.6643  1581.3191  1585.6340 32264.3747         NA         NA 
##     exGAUS        BCT 
##         NA         NA

## GAIG with k= 6.19 
## GAMLSS-RS iteration 1: Global Deviance = -874.5811 
## GAMLSS-RS iteration 2: Global Deviance = -887.0473 
## GAMLSS-RS iteration 3: Global Deviance = -888.7151 
## GAMLSS-RS iteration 4: Global Deviance = -888.9001 
## GAMLSS-RS iteration 5: Global Deviance = -888.9231 
## GAMLSS-RS iteration 6: Global Deviance = -888.9266 
## GAMLSS-RS iteration 7: Global Deviance = -888.9272

## GAIG with k= 6.19 
## ******************************************************************
## Family:  c("GIG", "Generalised Inverse Gaussian") 
## 
## Call:  gamlss(formula = CPUEw ~ factor(TGShort) + factor(survey) +  
##     factor(gear), family = names(getOrder(t1, 3)[1]),  
##     data = na.omit(subset(c4, CPUEw > 0))) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  log
## Mu Coefficients:
##                        Estimate Std. Error t value Pr(>|t|)    
## (Intercept)            -2.04502    0.15487 -13.205  < 2e-16 ***
## factor(TGShort)Coral.  -1.74291    1.01966  -1.709 0.088045 .  
## factor(TGShort)Herb.   -0.48068    0.25446  -1.889 0.059493 .  
## factor(TGShort)Invert. -0.36948    0.11450  -3.227 0.001338 ** 
## factor(TGShort)Plankt. -0.42440    0.35379  -1.200 0.230885    
## factor(survey)2013002   0.01996    0.11177   0.179 0.858315    
## factor(gear)TB          0.49349    0.14542   3.394 0.000747 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.47835    0.04813   9.939   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Nu link function:  identity 
## Nu Coefficients:
##             Estimate Std. Error t value Pr(>|t|)  
## (Intercept)  -0.2716     0.1427  -1.904   0.0575 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## No. of observations in the fit:  487 
## Degrees of Freedom for the fit:  9
##       Residual Deg. of Freedom:  478 
##                       at cycle:  7 
##  
## Global Deviance:     -888.9272 
##             AIC:     -870.9272 
##             SBC:     -833.2328 
## ******************************************************************

## GAIG with k= 6.19 
## GAIG with k= 6.19 
## GAIG with k= 6.19

## Single term deletions for
## mu
## 
## Model:
## CPUEw ~ factor(TGShort) + factor(survey) + factor(gear)
##                 Df     AIC     LRT   Pr(Chi)    
## <none>             -870.93                      
## factor(TGShort)  4 -861.74 17.1873 0.0017775 ** 
## factor(survey)   1 -872.91  0.0215 0.8834256    
## factor(gear)     1 -861.64 11.2904 0.0007791 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## 
## Family:  c("GIG", "Generalised Inverse Gaussian") 
## Fitting method: "nlminb" 
## 
## Call:  gamlssML(formula = CPUEw, family = "GIG", data = subset(c4,  
##     CPUEw > 0)) 
## 
## Mu Coefficients:
## [1]  -1.746
## Sigma Coefficients:
## [1]  0.5548
## Nu Coefficients:
## [1]  -0.02033
## 
##  Degrees of Freedom for the fit: 3 Residual Deg. of Freedom   756 
## Global Deviance:     -1304.45 
##             AIC:     -1298.45 
##             SBC:     -1284.56

Gear and trophic groups were significant variables in the model when the smoothing funciton was dropped. In the full model gears, the Nov. 2012 and Nov. 2005 surveys as well as the carnivores (intercept), coralivors, herbivores and invertivores were all significant factors affecting the catch rates (CPUEw) in the GAM model of CPUEw by trophic group, survey and gear (Generalised Inverse Gaussian distribution, AIC: -870, df: 9).

Gear had a highly significant impact on the catch composition, and all tropic groups except planktivores had a significant effect on catches. This is different from the PCA results where planktivores were a very clear factor explaining the variability, but they were caught in just 14 stations.

Conclusion from PCA and GAM model:
* Gear significantly affects catch rates
* Trophic groups (except planktivores) significantly affect catch rates
* Both Nov surveys (2012 and 2013) had significant effect on catches

---

Test differences in CPUE By Species in trap catches, by range of traits and factors

Use GAM model and evaluate how CPUE by weight (CPUEw) was affected by physical / area /surve factors and the traits : - depth - survey (random factor) - area - family group - trophic group - trophic level (linked to trophic group) - Place in water column - Diel activity - Habitat - Gregariousness - MaxLength

## GAMLSS-RS iteration 1: Global Deviance = -547.2298 
## GAMLSS-RS iteration 2: Global Deviance = -547.2336 
## GAMLSS-RS iteration 3: Global Deviance = -547.2336

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = fv)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in `formals<-`(`*tmp*`, envir = new.env(), value = alist(nu = nu)):
## 'fun' is not a function

## Warning in body(fun): argument is not a function

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in RS(): Algorithm RS has not yet converged

## Warning in additive.fit(x = X, y = wv, w = wt * w, s = s, who = who,
## smooth.frame, : additive.fit convergence not obtained in 30 iterations

## Warning in RS(): Algorithm RS has not yet converged

## minimum GAIC(k= 2 ) family: BCPEo 
## minimum GAIC(k= 3.84 ) family: BCPEo 
## minimum GAIC(k= 5.95 ) family: BCPEo

## GAIG with k= 5.95

##      BCPEo         GG        BCT      BCCGo       BCTo        GB2     IGAMMA 
## -1492.6272 -1436.5471 -1435.7982 -1432.0924 -1426.1289 -1412.4678 -1340.8260 
##        GIG         IG      LOGNO        LNO         GA        WEI       WEI3 
## -1334.8753 -1311.4019 -1272.3433 -1272.3433 -1182.2330 -1120.5560 -1120.5555 
##       WEI2        EXP    PARETO2         GP     exGAUS   PARETO2o     LOGNO2 
## -1120.4744  -987.9108  -966.8436  -954.7179  -791.9375  -157.3380         NA 
##       BCCG       BCPE 
##         NA         NA

## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = -1053.862 
## GAMLSS-RS iteration 2: Global Deviance = -1272.829 
## GAMLSS-RS iteration 3: Global Deviance = -1478.911 
## GAMLSS-RS iteration 4: Global Deviance = -1567.273 
## GAMLSS-RS iteration 5: Global Deviance = -1596.976 
## GAMLSS-RS iteration 6: Global Deviance = -1611.556 
## GAMLSS-RS iteration 7: Global Deviance = -1621.143 
## GAMLSS-RS iteration 8: Global Deviance = -1627.736 
## GAMLSS-RS iteration 9: Global Deviance = -1631.88 
## GAMLSS-RS iteration 10: Global Deviance = -1637.184 
## GAMLSS-RS iteration 11: Global Deviance = -1641.04 
## GAMLSS-RS iteration 12: Global Deviance = -1644.457 
## GAMLSS-RS iteration 13: Global Deviance = -1645.356 
## GAMLSS-RS iteration 14: Global Deviance = -1647.696 
## GAMLSS-RS iteration 15: Global Deviance = -1648.632 
## GAMLSS-RS iteration 16: Global Deviance = -1649.688 
## GAMLSS-RS iteration 17: Global Deviance = -1657.761 
## GAMLSS-RS iteration 18: Global Deviance = -1650.215 
## GAMLSS-RS iteration 19: Global Deviance = -1650.324 
## GAMLSS-RS iteration 20: Global Deviance = -1653.277

## Warning in RS(): Algorithm RS has not yet converged

## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = -1060.41 
## GAMLSS-RS iteration 2: Global Deviance = -1294.598 
## GAMLSS-RS iteration 3: Global Deviance = -1492.83 
## GAMLSS-RS iteration 4: Global Deviance = -1571.542 
## GAMLSS-RS iteration 5: Global Deviance = -1595.533 
## GAMLSS-RS iteration 6: Global Deviance = -1612.12 
## GAMLSS-RS iteration 7: Global Deviance = -1620.746 
## GAMLSS-RS iteration 8: Global Deviance = -1626.38 
## GAMLSS-RS iteration 9: Global Deviance = -1635.735 
## GAMLSS-RS iteration 10: Global Deviance = -1641.23 
## GAMLSS-RS iteration 11: Global Deviance = -1646.426 
## GAMLSS-RS iteration 12: Global Deviance = -1650.153 
## GAMLSS-RS iteration 13: Global Deviance = -1651.663 
## GAMLSS-RS iteration 14: Global Deviance = -1656.905 
## GAMLSS-RS iteration 15: Global Deviance = -1656.113 
## GAMLSS-RS iteration 16: Global Deviance = -1654.784 
## GAMLSS-RS iteration 17: Global Deviance = -1650.86 
## GAMLSS-RS iteration 18: Global Deviance = -1663.524 
## GAMLSS-RS iteration 19: Global Deviance = -1661.488 
## GAMLSS-RS iteration 20: Global Deviance = -1660.287

## Warning in RS(): Algorithm RS has not yet converged

## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = -1035.345 
## GAMLSS-RS iteration 2: Global Deviance = -1280.977 
## GAMLSS-RS iteration 3: Global Deviance = -1484.871 
## GAMLSS-RS iteration 4: Global Deviance = -1557.985 
## GAMLSS-RS iteration 5: Global Deviance = -1574.311 
## GAMLSS-RS iteration 6: Global Deviance = -1584.876 
## GAMLSS-RS iteration 7: Global Deviance = -1590.801 
## GAMLSS-RS iteration 8: Global Deviance = -1594.055 
## GAMLSS-RS iteration 9: Global Deviance = -1595.528 
## GAMLSS-RS iteration 10: Global Deviance = -1597.01 
## GAMLSS-RS iteration 11: Global Deviance = -1598.543 
## GAMLSS-RS iteration 12: Global Deviance = -1599.488 
## GAMLSS-RS iteration 13: Global Deviance = -1600.43 
## GAMLSS-RS iteration 14: Global Deviance = -1600.269 
## GAMLSS-RS iteration 15: Global Deviance = -1601.414 
## GAMLSS-RS iteration 16: Global Deviance = -1602.015 
## GAMLSS-RS iteration 17: Global Deviance = -1602.194 
## GAMLSS-RS iteration 18: Global Deviance = -1602.506 
## GAMLSS-RS iteration 19: Global Deviance = -1602.783 
## GAMLSS-RS iteration 20: Global Deviance = -1603.017

## Warning in RS(): Algorithm RS has not yet converged

## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = -1049.581 
## GAMLSS-RS iteration 2: Global Deviance = -1272.852 
## GAMLSS-RS iteration 3: Global Deviance = -1454.023 
## GAMLSS-RS iteration 4: Global Deviance = -1553.839 
## GAMLSS-RS iteration 5: Global Deviance = -1578.467 
## GAMLSS-RS iteration 6: Global Deviance = -1592.421 
## GAMLSS-RS iteration 7: Global Deviance = -1605.87 
## GAMLSS-RS iteration 8: Global Deviance = -1613.033 
## GAMLSS-RS iteration 9: Global Deviance = -1618.397 
## GAMLSS-RS iteration 10: Global Deviance = -1623.032 
## GAMLSS-RS iteration 11: Global Deviance = -1627.04 
## GAMLSS-RS iteration 12: Global Deviance = -1627.661 
## GAMLSS-RS iteration 13: Global Deviance = -1629.317 
## GAMLSS-RS iteration 14: Global Deviance = -1634.589 
## GAMLSS-RS iteration 15: Global Deviance = -1632.618 
## GAMLSS-RS iteration 16: Global Deviance = -1636.721 
## GAMLSS-RS iteration 17: Global Deviance = -1633.229 
## GAMLSS-RS iteration 18: Global Deviance = -1626.397 
## GAMLSS-RS iteration 19: Global Deviance = -1632.096 
## GAMLSS-RS iteration 20: Global Deviance = -1640.729

## Warning in RS(): Algorithm RS has not yet converged

## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = -849.2565 
## GAMLSS-RS iteration 2: Global Deviance = -1307.536 
## GAMLSS-RS iteration 3: Global Deviance = -1492.147 
## GAMLSS-RS iteration 4: Global Deviance = -1549.361 
## GAMLSS-RS iteration 5: Global Deviance = -1563.83 
## GAMLSS-RS iteration 6: Global Deviance = -1573.524 
## GAMLSS-RS iteration 7: Global Deviance = -1577.514 
## GAMLSS-RS iteration 8: Global Deviance = -1580.043 
## GAMLSS-RS iteration 9: Global Deviance = -1582.05 
## GAMLSS-RS iteration 10: Global Deviance = -1583.184 
## GAMLSS-RS iteration 11: Global Deviance = -1584.119 
## GAMLSS-RS iteration 12: Global Deviance = -1584.713 
## GAMLSS-RS iteration 13: Global Deviance = -1585.032 
## GAMLSS-RS iteration 14: Global Deviance = -1585.618 
## GAMLSS-RS iteration 15: Global Deviance = -1585.938 
## GAMLSS-RS iteration 16: Global Deviance = -1586.196 
## GAMLSS-RS iteration 17: Global Deviance = -1586.411 
## GAMLSS-RS iteration 18: Global Deviance = -1586.594 
## GAMLSS-RS iteration 19: Global Deviance = -1586.746 
## GAMLSS-RS iteration 20: Global Deviance = -1586.88

## Warning in RS(): Algorithm RS has not yet converged

## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = -909.9941 
## GAMLSS-RS iteration 2: Global Deviance = -1323.172 
## GAMLSS-RS iteration 3: Global Deviance = -1507.654 
## GAMLSS-RS iteration 4: Global Deviance = -1570.164 
## GAMLSS-RS iteration 5: Global Deviance = -1590.38 
## GAMLSS-RS iteration 6: Global Deviance = -1602.323 
## GAMLSS-RS iteration 7: Global Deviance = -1609.831 
## GAMLSS-RS iteration 8: Global Deviance = -1615.221 
## GAMLSS-RS iteration 9: Global Deviance = -1619.614 
## GAMLSS-RS iteration 10: Global Deviance = -1623.319 
## GAMLSS-RS iteration 11: Global Deviance = -1626.334 
## GAMLSS-RS iteration 12: Global Deviance = -1627.251 
## GAMLSS-RS iteration 13: Global Deviance = -1628.859 
## GAMLSS-RS iteration 14: Global Deviance = -1632.417 
## GAMLSS-RS iteration 15: Global Deviance = -1631.2 
## GAMLSS-RS iteration 16: Global Deviance = -1630.917 
## GAMLSS-RS iteration 17: Global Deviance = -1630.799 
## GAMLSS-RS iteration 18: Global Deviance = -1630.083 
## GAMLSS-RS iteration 19: Global Deviance = -1631.172 
## GAMLSS-RS iteration 20: Global Deviance = -1636.033

## Warning in RS(): Algorithm RS has not yet converged

## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = -852.4636 
## GAMLSS-RS iteration 2: Global Deviance = -1302.019 
## GAMLSS-RS iteration 3: Global Deviance = -1491.362 
## GAMLSS-RS iteration 4: Global Deviance = -1548.136 
## GAMLSS-RS iteration 5: Global Deviance = -1559.296 
## GAMLSS-RS iteration 6: Global Deviance = -1565.807 
## GAMLSS-RS iteration 7: Global Deviance = -1569.99 
## GAMLSS-RS iteration 8: Global Deviance = -1571.893 
## GAMLSS-RS iteration 9: Global Deviance = -1573.047 
## GAMLSS-RS iteration 10: Global Deviance = -1573.808 
## GAMLSS-RS iteration 11: Global Deviance = -1574.241 
## GAMLSS-RS iteration 12: Global Deviance = -1574.694 
## GAMLSS-RS iteration 13: Global Deviance = -1574.971 
## GAMLSS-RS iteration 14: Global Deviance = -1575.176 
## GAMLSS-RS iteration 15: Global Deviance = -1575.338 
## GAMLSS-RS iteration 16: Global Deviance = -1575.457 
## GAMLSS-RS iteration 17: Global Deviance = -1575.558 
## GAMLSS-RS iteration 18: Global Deviance = -1575.631 
## GAMLSS-RS iteration 19: Global Deviance = -1575.692 
## GAMLSS-RS iteration 20: Global Deviance = -1575.743

## Warning in RS(): Algorithm RS has not yet converged

##       df       AIC
## mod8  26 -1608.287
## mod7  27 -1599.277
## mod10 25 -1590.729
## mod12 24 -1588.033
## mod9  20 -1563.018
## mod11 19 -1548.880
## mod13 18 -1539.743

## GAIG with k= 5.95 
## ******************************************************************
## Family:  c("BCPEo", "Box-Cox Power Exponential-orig.") 
## 
## Call:  gamlss(formula = CPUEn ~ factor(id) + TrophicLevel +  
##     factor(TGShort) + factor(WaterCol) + factor(DielActivity) +  
##     factor(Habitat) + factor(Gregariousness) + MaxLength,  
##     family = names(getOrder(t1, 3)[1]), data = na.omit(subset(c4,  
##         gear == "TB" & CPUEn > 0))) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  log
## Mu Coefficients:
##                                             Estimate Std. Error  t value
## (Intercept)                               -2.640e+00  2.779e-03 -949.905
## factor(id)2                               -3.167e-02  2.211e-03  -14.323
## factor(id)3                               -5.397e-02  2.866e-03  -18.830
## factor(id)4                                1.769e-02  2.818e-03    6.276
## factor(id)5                               -6.388e-02  2.412e-03  -26.490
## factor(id)6                               -1.194e-01  2.473e-03  -48.282
## factor(id)7                               -7.969e-02  2.744e-03  -29.037
## TrophicLevel                               5.221e-03  2.136e-04   24.438
## factor(TGShort)Coral.                     -2.434e-01  1.486e-03 -163.809
## factor(TGShort)Herb.                       3.930e-01  1.746e-03  225.084
## factor(TGShort)Invert.                    -7.463e-02  1.618e-03  -46.133
## factor(TGShort)Plankt.                     3.163e+00  2.690e-03 1175.863
## factor(WaterCol)Demersal                  -7.242e-02  1.699e-03  -42.621
## factor(WaterCol)pelagic non-site attached  1.358e-01  2.369e-03   57.327
## factor(WaterCol)pelagic site attached     -2.616e+00  3.052e-03 -857.319
## factor(DielActivity)Night                  7.333e-02  1.788e-03   41.013
## factor(Habitat)Coral                       2.622e+00  3.363e-03  779.706
## factor(Habitat)sand                        4.961e-01  2.394e-03  207.193
## factor(Gregariousness)2                    2.375e-02  9.303e-04   25.524
## factor(Gregariousness)3                   -2.452e-01  2.187e-03 -112.085
## MaxLength                                 -4.699e-04  9.235e-05   -5.087
##                                           Pr(>|t|)    
## (Intercept)                                < 2e-16 ***
## factor(id)2                                < 2e-16 ***
## factor(id)3                                < 2e-16 ***
## factor(id)4                               9.97e-10 ***
## factor(id)5                                < 2e-16 ***
## factor(id)6                                < 2e-16 ***
## factor(id)7                                < 2e-16 ***
## TrophicLevel                               < 2e-16 ***
## factor(TGShort)Coral.                      < 2e-16 ***
## factor(TGShort)Herb.                       < 2e-16 ***
## factor(TGShort)Invert.                     < 2e-16 ***
## factor(TGShort)Plankt.                     < 2e-16 ***
## factor(WaterCol)Demersal                   < 2e-16 ***
## factor(WaterCol)pelagic non-site attached  < 2e-16 ***
## factor(WaterCol)pelagic site attached      < 2e-16 ***
## factor(DielActivity)Night                  < 2e-16 ***
## factor(Habitat)Coral                       < 2e-16 ***
## factor(Habitat)sand                        < 2e-16 ***
## factor(Gregariousness)2                    < 2e-16 ***
## factor(Gregariousness)3                    < 2e-16 ***
## MaxLength                                 5.86e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)   
## (Intercept)   1.0778     0.3316   3.251  0.00126 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Nu link function:  identity 
## Nu Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  -0.8592     0.1745  -4.923  1.3e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Tau link function:  log 
## Tau Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   -1.543      0.108  -14.29   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## No. of observations in the fit:  384 
## Degrees of Freedom for the fit:  24
##       Residual Deg. of Freedom:  360 
##                       at cycle:  20 
##  
## Global Deviance:     -1636.033 
##             AIC:     -1588.033 
##             SBC:     -1493.218 
## ******************************************************************

Evaluting catches of fish in relation to their traits. Should be evaluated by numbers, for single large fish not to dominate the results. The model based on CPUE by numbers (mod12, where surveys and depth are excluded) had lower AIC than models with numbers and fishing hours as a factor. Also, including survey as an ordinary factor led to lower AIC than including it as a random factor.

Model (BCPOe Box-Cox Power Exponential-orig. distribution, AIC: -1619.583, df:24) results show that numbers of fish caught (CPUE by numbers) was significantly affected by:
* all areas * all traits, except MaxLength and WaterColumn (demersal)

Interpretation of results: If depth was included it was not a significant factor. CPUEn varies significantly by areas and by traits, and with the variability being consistent this indicates regional differences in catches of different trophic groups / traits /species.

CPUE by trophic group by area & by survey (year)

FIG 5 in Feb 2019 version of MS

Presents the CPUE pr hr, for each management area and survey, standardized by the number of traps set in each area X survey combination.

Trap catches were dominated by carnivores, occuring in all area X survey combinations, followed by invertivores who also occurred everywhere, albeight in lower densities than the carnivores.

Cathces of planktivores and invertivores varied substantially between areas and surveys, in contrast to carnivores who were caught in all surveys and areas. This explains why invertivores and planktivores were signifant factors on the catches (GAM model of CPUEn with traits), while carnivores were not.

Area 2, 5 and 6 seem to have the highest diversity of trophic groups, but this has to be calculated as functional diversity, using the FD package (or similar).

## Saving 7 x 5 in image

## quartz_off_screen 
##                 2

Trophic group CPUE by depth

A modified version of Figure 6 in Feb 2019 MS.

Basically a more complex version of the previous plot. Adds the depth dimension to the plot, and the surveys as colours.

Shows uniform catch rates of carnivores and invertivores from 0-50m.

All surveys seem to overlap, so no difference in trophich group depth dependent catch rates between the surveys.

## quartz_off_screen 
##                 2

Trophic level in Trap catches

## quartz_off_screen 
##                 2

GAM model of trophic level of catches

## GAMLSS-RS iteration 1: Global Deviance = 469.1735 
## GAMLSS-RS iteration 2: Global Deviance = 469.1735

## minimum GAIC(k= 2 ) family: BCPEo 
## minimum GAIC(k= 3.84 ) family: BCPEo 
## minimum GAIC(k= 6.01 ) family: BCPEo

## GAIG with k= 6.01

##     BCPEo      BCPE       GB2        GG     BCCGo      BCCG      BCTo       BCT 
##  379.8398  379.9088  389.9433  391.0956  392.9895  393.0466  398.8034  398.8592 
##       WEI      WEI3      WEI2    exGAUS        GA     LOGNO       LNO    LOGNO2 
##  419.0743  419.0743  444.8026  535.7938  601.9244  644.7756  644.7756  644.8770 
##        IG    IGAMMA       EXP   PARETO2  PARETO2o        GP       GIG 
##  651.5663  695.9278 1966.2102 2005.7372 2010.1027 2010.2496        NA

## GAIG with k= 5.95

## [1] "BCPEo"

## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = 319.1556 
## GAMLSS-RS iteration 2: Global Deviance = 316.9397 
## GAMLSS-RS iteration 3: Global Deviance = 316.7322 
## GAMLSS-RS iteration 4: Global Deviance = 316.5108 
## GAMLSS-RS iteration 5: Global Deviance = 316.2424 
## GAMLSS-RS iteration 6: Global Deviance = 315.9263 
## GAMLSS-RS iteration 7: Global Deviance = 315.5498 
## GAMLSS-RS iteration 8: Global Deviance = 315.0915 
## GAMLSS-RS iteration 9: Global Deviance = 314.4959 
## GAMLSS-RS iteration 10: Global Deviance = 313.672 
## GAMLSS-RS iteration 11: Global Deviance = 312.4064 
## GAMLSS-RS iteration 12: Global Deviance = 310.4638 
## GAMLSS-RS iteration 13: Global Deviance = 308.8748 
## GAMLSS-RS iteration 14: Global Deviance = 308.1439 
## GAMLSS-RS iteration 15: Global Deviance = 307.8386 
## GAMLSS-RS iteration 16: Global Deviance = 307.7504 
## GAMLSS-RS iteration 17: Global Deviance = 307.7276 
## GAMLSS-RS iteration 18: Global Deviance = 307.7218 
## GAMLSS-RS iteration 19: Global Deviance = 307.7202 
## GAMLSS-RS iteration 20: Global Deviance = 307.7198

## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = 319.2872 
## GAMLSS-RS iteration 2: Global Deviance = 317.0106 
## GAMLSS-RS iteration 3: Global Deviance = 316.8008 
## GAMLSS-RS iteration 4: Global Deviance = 316.5776 
## GAMLSS-RS iteration 5: Global Deviance = 316.3096 
## GAMLSS-RS iteration 6: Global Deviance = 315.9928 
## GAMLSS-RS iteration 7: Global Deviance = 315.615 
## GAMLSS-RS iteration 8: Global Deviance = 315.1546 
## GAMLSS-RS iteration 9: Global Deviance = 314.5501 
## GAMLSS-RS iteration 10: Global Deviance = 313.7191 
## GAMLSS-RS iteration 11: Global Deviance = 312.4365 
## GAMLSS-RS iteration 12: Global Deviance = 310.468 
## GAMLSS-RS iteration 13: Global Deviance = 308.8815 
## GAMLSS-RS iteration 14: Global Deviance = 308.1497 
## GAMLSS-RS iteration 15: Global Deviance = 307.8475 
## GAMLSS-RS iteration 16: Global Deviance = 307.7486 
## GAMLSS-RS iteration 17: Global Deviance = 307.7384 
## GAMLSS-RS iteration 18: Global Deviance = 307.7218 
## GAMLSS-RS iteration 19: Global Deviance = 307.7211

## GAIG with k= 5.95 
## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = 319.1556 
## GAMLSS-RS iteration 2: Global Deviance = 316.9397 
## GAMLSS-RS iteration 3: Global Deviance = 316.7322 
## GAMLSS-RS iteration 4: Global Deviance = 316.5108 
## GAMLSS-RS iteration 5: Global Deviance = 316.2424 
## GAMLSS-RS iteration 6: Global Deviance = 315.9263 
## GAMLSS-RS iteration 7: Global Deviance = 315.5498 
## GAMLSS-RS iteration 8: Global Deviance = 315.0915 
## GAMLSS-RS iteration 9: Global Deviance = 314.4959 
## GAMLSS-RS iteration 10: Global Deviance = 313.672 
## GAMLSS-RS iteration 11: Global Deviance = 312.4064 
## GAMLSS-RS iteration 12: Global Deviance = 310.4638 
## GAMLSS-RS iteration 13: Global Deviance = 308.8748 
## GAMLSS-RS iteration 14: Global Deviance = 308.1439 
## GAMLSS-RS iteration 15: Global Deviance = 307.8386 
## GAMLSS-RS iteration 16: Global Deviance = 307.7504 
## GAMLSS-RS iteration 17: Global Deviance = 307.7276 
## GAMLSS-RS iteration 18: Global Deviance = 307.7218 
## GAMLSS-RS iteration 19: Global Deviance = 307.7202 
## GAMLSS-RS iteration 20: Global Deviance = 307.7198 
## GAMLSS-RS iteration 1: Global Deviance = 315.7557 
## GAMLSS-RS iteration 2: Global Deviance = 314.4354 
## GAMLSS-RS iteration 3: Global Deviance = 314.0963 
## GAMLSS-RS iteration 4: Global Deviance = 313.7754 
## GAMLSS-RS iteration 5: Global Deviance = 313.3731 
## GAMLSS-RS iteration 6: Global Deviance = 312.8346 
## GAMLSS-RS iteration 7: Global Deviance = 311.2386 
## GAMLSS-RS iteration 8: Global Deviance = 307.5397 
## GAMLSS-RS iteration 9: Global Deviance = 304.4499 
## GAMLSS-RS iteration 10: Global Deviance = 302.674 
## GAMLSS-RS iteration 11: Global Deviance = 301.713 
## GAMLSS-RS iteration 12: Global Deviance = 301.2475 
## GAMLSS-RS iteration 13: Global Deviance = 300.9538 
## GAMLSS-RS iteration 14: Global Deviance = 300.6442 
## GAMLSS-RS iteration 15: Global Deviance = 300.3321 
## GAMLSS-RS iteration 16: Global Deviance = 300.0835 
## GAMLSS-RS iteration 17: Global Deviance = 299.9489 
## GAMLSS-RS iteration 18: Global Deviance = 299.8841 
## GAMLSS-RS iteration 19: Global Deviance = 299.8577 
## GAMLSS-RS iteration 20: Global Deviance = 299.8512

## GAIG with k= 5.95 
## GAIG with k= 5.95 
## GAMLSS-RS iteration 1: Global Deviance = 319.1556 
## GAMLSS-RS iteration 2: Global Deviance = 316.9397 
## GAMLSS-RS iteration 3: Global Deviance = 316.7322 
## GAMLSS-RS iteration 4: Global Deviance = 316.5108 
## GAMLSS-RS iteration 5: Global Deviance = 316.2424 
## GAMLSS-RS iteration 6: Global Deviance = 315.9263 
## GAMLSS-RS iteration 7: Global Deviance = 315.5498 
## GAMLSS-RS iteration 8: Global Deviance = 315.0915 
## GAMLSS-RS iteration 9: Global Deviance = 314.4959 
## GAMLSS-RS iteration 10: Global Deviance = 313.672 
## GAMLSS-RS iteration 11: Global Deviance = 312.4064 
## GAMLSS-RS iteration 12: Global Deviance = 310.4638 
## GAMLSS-RS iteration 13: Global Deviance = 308.8748 
## GAMLSS-RS iteration 14: Global Deviance = 308.1439 
## GAMLSS-RS iteration 15: Global Deviance = 307.8386 
## GAMLSS-RS iteration 16: Global Deviance = 307.7504 
## GAMLSS-RS iteration 17: Global Deviance = 307.7276 
## GAMLSS-RS iteration 18: Global Deviance = 307.7218 
## GAMLSS-RS iteration 19: Global Deviance = 307.7202 
## GAMLSS-RS iteration 20: Global Deviance = 307.7198 
## GAMLSS-RS iteration 1: Global Deviance = 315.9843 
## GAMLSS-RS iteration 2: Global Deviance = 314.4932 
## GAMLSS-RS iteration 3: Global Deviance = 314.0544 
## GAMLSS-RS iteration 4: Global Deviance = 313.6932 
## GAMLSS-RS iteration 5: Global Deviance = 313.2687 
## GAMLSS-RS iteration 6: Global Deviance = 312.167 
## GAMLSS-RS iteration 7: Global Deviance = 309.4126 
## GAMLSS-RS iteration 8: Global Deviance = 305.8132 
## GAMLSS-RS iteration 9: Global Deviance = 303.4394 
## GAMLSS-RS iteration 10: Global Deviance = 302.0941 
## GAMLSS-RS iteration 11: Global Deviance = 301.4134 
## GAMLSS-RS iteration 12: Global Deviance = 301.051 
## GAMLSS-RS iteration 13: Global Deviance = 300.7309 
## GAMLSS-RS iteration 14: Global Deviance = 300.3878 
## GAMLSS-RS iteration 15: Global Deviance = 300.1884 
## GAMLSS-RS iteration 16: Global Deviance = 300.0751 
## GAMLSS-RS iteration 17: Global Deviance = 300.0355 
## GAMLSS-RS iteration 18: Global Deviance = 300.0217 
## GAMLSS-RS iteration 19: Global Deviance = 300.0187 
## GAMLSS-RS iteration 20: Global Deviance = 300.0123

##           df      AIC
## tl1 12.00000 331.7198
## tl2 12.99961 333.7203
## tl4 19.08503 338.1823
## tl3 20.00000 339.8512

## GAIG with k= 5.95 
## ******************************************************************
## Family:  c("BCPEo", "Box-Cox Power Exponential-orig.") 
## 
## Call:  gamlss(formula = TrophicLevel ~ depth + factor(id) +  
##     factor(survey), family = names(getOrder(t1, 3)[1]),  
##     data = na.omit(subset(catch.traits, gear == "TB" &  
##         TrophicLevel != "NA"))) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  log
## Mu Coefficients:
##                         Estimate Std. Error t value Pr(>|t|)    
## (Intercept)            1.3633473  0.0135997 100.249   <2e-16 ***
## depth                 -0.0001769  0.0001835  -0.965   0.3354    
## factor(id)2            0.0036357  0.0110590   0.329   0.7425    
## factor(id)3            0.0301380  0.0146015   2.064   0.0397 *  
## factor(id)4            0.0034203  0.0237053   0.144   0.8853    
## factor(id)5            0.0026418  0.0136358   0.194   0.8465    
## factor(id)6           -0.0043772  0.0117893  -0.371   0.7106    
## factor(id)7            0.0034907  0.0124887   0.280   0.7800    
## factor(survey)2013002  0.0007028  0.0063718   0.110   0.9122    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) -2.47860    0.04204  -58.96   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Nu link function:  identity 
## Nu Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   5.4618     0.4472   12.21   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Tau link function:  log 
## Tau Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.15467    0.09677   11.93   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## No. of observations in the fit:  408 
## Degrees of Freedom for the fit:  12
##       Residual Deg. of Freedom:  396 
##                       at cycle:  20 
##  
## Global Deviance:     307.7198 
##             AIC:     331.7198 
##             SBC:     379.8551 
## ******************************************************************

The GAM Model (Box-Cox Power Exponential, AIC: 331.72, df: 12) of trophic level by depth, area and surveys only showed area 3 and the intercept to be significant factors affecting the Trophic level of catches meaning that trophic level was similar in all areas except area 3 (and 1)

Maxlength in catches

## Saving 7 x 5 in image

## quartz_off_screen 
##                 2

Place in water column

## Saving 7 x 5 in image

## quartz_off_screen 
##                 2

Diel activity

## Saving 7 x 5 in image

## quartz_off_screen 
##                 2

Gregariousness

## Saving 7 x 5 in image

## quartz_off_screen 
##                 2

---

Biodiversity (species densities, trophic diversity etc)

Aggregated Species density per area for (traps only)

Species per area pr hour of fishing, traps only.

Area	Nov.2012	May 2013	Nov.2013	Average
1	0.0085068	0.0178680	0.0190901	0.0132567
2	0.0115513	0.0129490	0.0110088	0.0120432
3	0.0173986	0.0239815	0.0185053	0.0210644
4	0.0259457	0.0253077	0.0214874	0.0235929
5	0.0129828	0.0170187	0.0113921	0.0138271
6	0.0133471	0.0131170	0.0129926	0.0131638
7	0.0138589	0.0128944	0.0149727	0.0136772

The highest number of species was observed in area 2 (36), followed by area 3, 6 and 7 (25 in each).

Average species density across the three surveys(number of species pr hour of fishing - soak time of traps) ranged from 0.01204 (area 2) to 0.02359 (area 4). Area 4 had the highest species density across all surveys (table above), while the area with the lowest species density varied by survey: area 1 in November 2012 , area 7 in May 2013 , and area 2 in November 2013.

Species density pr station - Zero Adjusted GAM model

## GAMLSS-RS iteration 1: Global Deviance = -1013.13 
## GAMLSS-RS iteration 2: Global Deviance = -1013.13

## GAMLSS-RS iteration 1: Global Deviance = -1013.112 
## GAMLSS-RS iteration 2: Global Deviance = -1013.108 
## GAMLSS-RS iteration 3: Global Deviance = -1013.108

## GAMLSS-RS iteration 1: Global Deviance = -1057.187 
## GAMLSS-RS iteration 2: Global Deviance = -1057.187

## GAMLSS-RS iteration 1: Global Deviance = -1056.858 
## GAMLSS-RS iteration 2: Global Deviance = -1056.854 
## GAMLSS-RS iteration 3: Global Deviance = -1056.854

## GAMLSS-RS iteration 1: Global Deviance = -986.5036 
## GAMLSS-RS iteration 2: Global Deviance = -986.5036

##               df        AIC
## mod.sd3 21.00000 -1015.1866
## mod.sd4 22.68257 -1011.4893
## mod.sd1 12.00000  -989.1301
## mod.sd2 12.95091  -987.2061
## mod.sd5 21.00000  -944.5036

## ******************************************************************
## Family:  c("ZAGA", "Zero Adjusted GA") 
## 
## Call:  gamlss(formula = Species_Hrs ~ depth + factor(area) +  
##     factor(survey), nu.formula = Species_Hrs ~ depth +  
##     factor(area) + factor(survey), family = ZAGA, data = spue.df) 
## 
## Fitting method: RS() 
## 
## ------------------------------------------------------------------
## Mu link function:  log
## Mu Coefficients:
##                         Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -3.053e+00  4.470e-02 -68.310  < 2e-16 ***
## depth                  5.284e-05  5.820e-04   0.091  0.92769    
## factor(area)2          1.274e-01  3.987e-02   3.194  0.00147 ** 
## factor(area)3          1.072e-01  4.995e-02   2.147  0.03216 *  
## factor(area)4          1.324e-01  6.213e-02   2.130  0.03351 *  
## factor(area)5          1.217e-01  4.869e-02   2.500  0.01268 *  
## factor(area)6         -2.721e-03  4.122e-02  -0.066  0.94739    
## factor(area)7         -3.028e-02  4.695e-02  -0.645  0.51922    
## factor(survey)2013002  1.999e-01  2.572e-02   7.774 2.96e-14 ***
## factor(survey)2013005  2.402e-01  3.285e-02   7.310 7.81e-13 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Sigma link function:  log
## Sigma Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) -1.60349    0.03933  -40.77   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## Nu link function:  logit 
## Nu Coefficients:
##                        Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -0.138533   0.307058  -0.451   0.6520    
## depth                 -0.002958   0.004342  -0.681   0.4959    
## factor(area)2          0.109462   0.281142   0.389   0.6971    
## factor(area)3          0.258227   0.346525   0.745   0.4564    
## factor(area)4          0.024677   0.429925   0.057   0.9542    
## factor(area)5          0.396141   0.327785   1.209   0.2273    
## factor(area)6         -0.521334   0.306163  -1.703   0.0891 .  
## factor(area)7          0.326295   0.322112   1.013   0.3114    
## factor(survey)2013002 -0.010748   0.188148  -0.057   0.9545    
## factor(survey)2013005  0.994981   0.213500   4.660 3.83e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## ------------------------------------------------------------------
## No. of observations in the fit:  674 
## Degrees of Freedom for the fit:  21
##       Residual Deg. of Freedom:  653 
##                       at cycle:  2 
##  
## Global Deviance:     -1057.187 
##             AIC:     -1015.187 
##             SBC:     -920.4088 
## ******************************************************************

The variability and relative difference in rank of species density per area and survey as observed in the table above was confirmed by the zero-adjusted GAMLSS model with specified nu-function (with “ZAGA” distribution, AIC: -1015, df: 21) where all surveys and areas 1, 2, 3, 4, and 5 (but not depth) were significant factors in the model.

---

Functional diversity

We used the same approach as Stuart-Smith et al to exclude very few observations of a single species and excluded all species observations occuring in 2 or fewer stations (traps).

The ‘dbFD’ function only works when run in-line, not the whole code-chunk.

To avoid dbFD crashing ‘m’ was set in the ‘dbFD’ call (the number of PoCA axes kept as traits to do the FRic analysis). Several levels of ‘m’ were tested, and this could probably be tested further (to higher levels of ‘m’)

Results for n_spec > 2, m=10

Results based on the output from the dbFD analysis:

Overall Fric (R2)= 0.6868565

No area came out with the highest or lowest FD across all four metrics, but some patterns emerged:

• Area 1 was highest in FEve and FDiv, but second lowest in FDis and RaoQ
• Area 2 was consistently high in all metrics, and highest in FDis and RaoQ
• Area 3 was above average in all metrics, except FDiv
• Area 4 was lowest in RaoQ and FDis, and below average in all others
• Area 5 was above average in all metrics
• Area 6 was above average in FEve, but below average in all others, and had the lowest value in FDis
• Area 7 was above average in RaoQ, had the lowest value in FEve, and below average in FDiv and FDis

Overall the highest Functional Diversity metrics were found in areas 1 and 2, while the lowest were found in areas 4, 6 and 7.

RaoQ is the most used metric (the one used by Stuart-Smith et al), and this is the metric we will use and discuss in the MS. It supports the analysis of species densities for Area 2, although Area 6 has the second lowest RaoQ which indicates that although the number of species is higher, their functional diversity is lower (or what we sampled was lower) compared to the other areas.

Stuart-Smith tests various RaoQs by taking out one trait at a time and this way ranking the traits according to R2.

## Species x species distance matrix was not Euclidean. Lingoes correction was applied. 
## FRic: Dimensionality reduction was required. The last 12 PCoA axes (out of 22 in total) were removed. 
## FRic: Quality of the reduced-space representation (based on corrected distance matrix) = 0.6868565

Area	FEve	FDiv	FDis	RaoQ
A1	0.86	0.792	0.237	0.065
A2	0.857	0.787	0.207	0.055
A3	0.808	0.812	0.248	0.077
A4	0.841	0.824	0.209	0.055
A5	0.849	0.785	0.229	0.066
A6	0.851	0.834	0.211	0.058
A7	0.813	0.84	0.18	0.038

Rerunning the FD model excluding one trait at a time

Need to reduce ‘m’ to 9 to avoid crash:

Zero distance(s)Error in convhulln(tr.FRic, “FA”) : Received error code 2 from qhull. Qhull error: QH6114 qhull precision error: initial simplex is not convex. Distance=1.4e-16

Removing the MaxLenght or Trophic level traits from the model lead to the largest increase of R2 (by 31% and 22% respectively) compared to the model with all traits included, and higher than removing any of the other traits from the model . R2 was reduced when removing any of the other five traits from the model, indicating that these traits should be kept in the model.

Reran the model excluding both MaxLength and Trophic Level:

Evaluation of runs: * Use the model without MaxLenght and Trophic Level * Removing Maxlength makes sence because the data included a few catches of sharks with very large MaxLengt (>2m) that skews the MaxLenght distribution. Excluding MaxLength creates an analysis around traits that are more common in distribution across species. + Also, the sharks are rowing predators, not so site attached (but that also applies to some pelagic species, especially the large ones, so the same argument applies for these - excluding large body size reduces the importance of these non-site attached species) * Removing TrophicLevel would be linked to MaxLength as the largest fish would be the apex predators (e.g. sharks)

## Species x species distance matrix was not Euclidean. Lingoes correction was applied. 
## FRic: Dimensionality reduction was required. The last 11 PCoA axes (out of 20 in total) were removed. 
## FRic: Quality of the reduced-space representation (based on corrected distance matrix) = 0.7082607 
## Species x species distance matrix was not Euclidean. Lingoes correction was applied. 
## FRic: Dimensionality reduction was required. The last 10 PCoA axes (out of 19 in total) were removed. 
## FRic: Quality of the reduced-space representation (based on corrected distance matrix) = 0.7018365 
## Species x species distance matrix was not Euclidean. Lingoes correction was applied. 
## FRic: Dimensionality reduction was required. The last 13 PCoA axes (out of 22 in total) were removed. 
## FRic: Quality of the reduced-space representation (based on corrected distance matrix) = 0.5803622 
## Species x species distance matrix was not Euclidean. Lingoes correction was applied. 
## FRic: Dimensionality reduction was required. The last 13 PCoA axes (out of 22 in total) were removed. 
## FRic: Quality of the reduced-space representation (based on corrected distance matrix) = 0.637926 
## Species x species distance matrix was not Euclidean. Lingoes correction was applied. 
## FRic: Dimensionality reduction was required. The last 13 PCoA axes (out of 22 in total) were removed. 
## FRic: Quality of the reduced-space representation (based on corrected distance matrix) = 0.6523521 
## Species x species distance matrix was not Euclidean. Lingoes correction was applied. 
## FRic: Dimensionality reduction was required. The last 13 PCoA axes (out of 22 in total) were removed. 
## FRic: Quality of the reduced-space representation (based on corrected distance matrix) = 0.6230467 
## Species x species distance matrix was not Euclidean. Lingoes correction was applied. 
## FRic: Dimensionality reduction was required. The last 13 PCoA axes (out of 22 in total) were removed. 
## FRic: Quality of the reduced-space representation (based on corrected distance matrix) = 0.6466908

## Species x species distance matrix was not Euclidean. Lingoes correction was applied. 
## FRic: Dimensionality reduction was required. The last 2 PCoA axes (out of 10 in total) were removed. 
## FRic: Quality of the reduced-space representation (based on corrected distance matrix) = 0.9327644

	All	MaxLength	Trophic.Level	Trophic.group	Water.column	Habitat	Gregariousness	diel	ML_TL
R2	0.687	0.708	0.702	0.58	0.638	0.652	0.623	0.647	0.933

	All	MaxLength	Trophic.Level	Trophic.group	Water.column	Habitat	Gregariousness	diel	ML_TL
A1	0.065	0.080	0.074	0.051	0.054	0.086	0.049	0.076	0.099
A2	0.055	0.066	0.062	0.038	0.060	0.069	0.048	0.058	0.082
A3	0.077	0.091	0.080	0.062	0.069	0.104	0.062	0.084	0.098
A4	0.055	0.074	0.061	0.045	0.047	0.070	0.040	0.064	0.093
A5	0.066	0.075	0.070	0.052	0.063	0.088	0.053	0.074	0.082
A6	0.058	0.067	0.055	0.040	0.062	0.078	0.054	0.068	0.066
A7	0.038	0.046	0.047	0.024	0.046	0.048	0.028	0.042	0.063

---

Results table

This table combines CPUE, traits pr gear type, species density and RaoQ into one results-table.

area	TrapCPUEw	GillnetCPUEw	TrapCPUEn	GillnetCPUEn	TrophicL_Traps	TrophicL_Gillnets	Greg_Traps	Greg_Gillnets	SpeciesDensity	RaoQ
1	0.117	1.639	0.060	0.675	3.825	3.933	1.510	2.412	0.013	0.099
2	0.149	1.176	0.086	1.366	3.815	3.891	1.478	2.261	0.012	0.082
3	0.136	0.309	0.075	0.268	3.844	4.140	1.553	2.500	0.021	0.098
4	0.059	0.698	0.067	1.001	3.765	4.284	1.357	2.107	0.024	0.093
5	0.079	0.726	0.090	1.237	3.797	3.853	1.627	2.176	0.014	0.082
6	0.135	0.270	0.072	0.768	3.762	3.927	1.531	2.286	0.013	0.066
7	0.098	0.683	0.075	1.235	3.828	3.612	1.565	1.900	0.014	0.063